CHAPTER 1 INTRODUCTION 1.

Introduction to Digital Image Processing:
   Vision allows humans to perceive and understand the world surrounding us. Computer vision aims to duplicate the effect of human vision by electronically perceiving and understanding an image. Giving computers the ability to see is not an easy task - we live in a three dimensional (3D) world, and when computers try to analyze objects in 3D space, available visual sensors (e.g., TV cameras) usually give two dimensional (2D) images, and this projection to a lower number of dimensions incurs an enormous loss of information.    In order to simplify the task of computer vision understanding, two levels are usually distinguished; low-level image processing and high level image understanding. Usually very little knowledge about the content of images High level processing is based on knowledge, goals, and plans of how to achieve those goals. Artificial intelligence (AI) methods are used in many cases. High-level computer vision tries to imitate human recognition and the ability to make decisions according to the information contained in the image.   This course deals almost exclusively with low-level image processing, high level in which is a continuation of this course. Age processing is discussed in the course Image Analysis and Understanding, which is a continuation of this course.

1.1 History:
Many of the techniques of digital image processing, or digital picture processing as it was often called, were developed in the 1960s at the Jet Propulsion Laboratory, MIT, Bell Labs, University of Maryland, and few other places, with application to satellite imagery, wire photo standards conversion, medical imaging, videophone, character recognition, and photo enhancement. But the cost of processing was fairly high with the computing equipment of that era. In the 1970s, digital image processing proliferated, when cheaper computers Creating a film or electronic image of any picture or paper form. It is accomplished by scanning or photographing an object and turning it into a matrix of dots (bitmap), the meaning of which is unknown to the computer, only to the human

viewer. Scanned images of text may be encoded into computer data (ASCII or EBCDIC) with page recognition software (OCR).

1.2 Basic Concepts:
 

A signal is a function depending on some variable with physical meaning. Signals can be
o o o o

One-dimensional (e.g., dependent on time), Two-dimensional (e.g., images dependent on two co-ordinates in a plane), Three-dimensional (e.g., describing an object in space), Or higher dimensional.

1.3 Pattern recognition
Pattern recognition is a field within the area of machine learning. Alternatively, it can be defined as "the act of taking in raw data and taking an action based on the category of the data" . As such, it is a collection of methods for supervised learning. Pattern recognition aims to classify data (patterns) based on either a priori knowledge or on statistical information extracted from the patterns. The patterns to be classified are usually groups of measurements or observations, defining points in an appropriate multidimensional space. Are to represent, for example, color images consisting of three component colors.

1.4 Image functions:
 

The image can be modeled by a continuous function of two or three variables; Arguments are co-ordinates x, y in a plane, while if images change in time a third variable t might be added.

 

The image function values correspond to the brightness at image points. The function value can express other physical quantities as well (temperature, pressure distribution, distance from the observer, etc.).

The brightness integrates different optical quantities - using brightness as a basic quantity allows us to avoid the description of the very complicated process of image formation.

The image on the human eye retina or on a TV camera sensor is intrinsically 2D. We shall call such a 2D image bearing information about brightness points an intensity image.

  

The real world, which surrounds us, is intrinsically 3D. The 2D intensity image is the result of a perspective projection of the 3D scene. When 3D objects are mapped into the camera plane by perspective projection a lot of information disappears as such a transformation is not one-to-one.

 

Recognizing or reconstructing objects in a 3D scene from one image is an ill-posed problem. Recovering information lost by perspective projection is only one, mainly geometric, problem of computer vision.

The second problem is how to understand image brightness. The only information available in an intensity image is brightness of the appropriate pixel, which is dependent on a number of independent factors such as
o

Object surface reflectance properties (given by the surface material, microstructure and marking),

o O

Illumination properties, And object surface orientation with respect to a viewer and light source.

CHAPTER 2 2. DIGITAL IMAGE FORENSICS
Today's technology allows digital media to be altered and manipulated in ways that were impossible twenty years ago. We are feeling the impact of this technology in nearly every corner of our lives, from the courts to the media, politics, business, and science. As this technology continues to evolve it will become increasingly more important for the science of digital forensics to keep pace. This presentation will describe state of the art techniques in digital image forensics. Digital watermarking has been proposed as a means by which an image can be authenticated. This approach works by inserting at the time of recording an imperceptible digital code (a watermark) into the image. With the assumption that tampering will alter a watermark, an image can be authenticated by verifying that the extracted watermark is the same as that which was inserted. The major drawback of this approach is that a watermark must be inserted at precisely the time of recording, which limits this approach to specially equipped digital cameras. In contrast, recent advances in digital forensics operate in the absence of any watermark or specialized hardware. With the assumption that tampering disturbs certain underlying statistical properties of an image, these forensic techniques can detect specific forms of tampering. Air-brushing or re-touching can be detected by measuring deviations of the underlying color filter array correlations. Specifically, virtually all digital cameras record only a subset of all the pixels needed for a full-resolution color image. Instead, only a subset of the pixels is recorded by a color filter array (CFA) placed atop the digital sensor. The most frequently used CFA, the Bayer array, employs three color filters: red, green, and blue. Since only a single color sample is recorded at each pixel location, the other two color samples must be estimated from the neighboring samples in order to obtain a three-channel color image. The estimation of the missing color samples is referred to as CFA interpolation or demosaicking. In its simplest form, the missing pixels are filled in by spatially averaging the recorded values. Since the CFA is arranged in a periodic pattern, a periodic set of pixels will be precisely correlated to their neighbors according to the CFA interpolation algorithm. When an image is re-touched, it is likely that these correlations will be destroyed. As such, the presence or lack of these correlations can be used to authenticate an image, or expose it as a forgery. A digital composite of two people can be detected by measuring differences in the direction to the illuminating light sources from their faces and body. By making some initial simplifying assumptions about the light and the surface being illuminated, we can mathematically express how much light a surface should receive as a function of its position relative to the light. A surface that is

For marking media files with copyright information. digital watermarks are only perceptible under certain conditions. This form of tampering can be detected by first partitioning an image into small blocks. The blocks are then re-ordered so that they are placed a distance to each other that is proportional to the differences in their pixel colors. i. Traditional Watermarks may be applied to visible media (like images or video). Digital watermarks may be used to verify the authenticity or integrity of the carrier signal or to show the identity of its owners. their presence can then be used as evidence of tampering. a region growing algorithm combines any significant number of neighboring blocks that are consistent with the cloning of an image region. Duplication or cloning is a simple and powerful form of manipulation used to remove objects or people from an image. With identical and highly similar blocks neighboring each other in the re-ordered sequence. video.directly facing the light. If a digital watermark distorts the carrier signal in a way that it becomes perceivable. Like traditional watermarks. The needed properties of a digital watermark depend on the use case in which it is applied. will be brighter than a surface that is turned away from the light. but does not need to contain a relation to the carrier signal. Instead. after using some algorithm.1. Any inconsistencies in lighting can then be used as evidence of tampering.e. the hidden information should. DIGITAL WATERMARKING A digital watermark is a kind of marker covertly embedded in a noise-tolerant signal such as audio or image data. a fragile watermark would be applied. Unlike metadata that is added to the carrier signal. . it is of no use. It is typically used to identify ownership of the copyright of such signal. Since it is statistically unlikely to find identical and spatially coherent regions in an image. for example. pictures. texts or 3D models. A signal may carry several different watermarks at the same time. whereas in digital watermarking. standard techniques can be used to determine the direction to the light source for any object or person in an image. It is prominently used for tracing copyright infringements and for banknote authentication. if integrity has to be ensured. a digital watermark does not change the size of the carrier signal. 2. the signal may be audio. "Watermarking" is the process of hiding digital information in a carrier signal. and imperceptible anytime else. Once expressed in this form. a digital watermark has to be rather robust against modifications that can be applied to the carrier signal.

this method is useless in the digital world. digital watermarking tries to control the robustness as top priority. digital watermarking is a passive protection tool. In particular. There are a variety of image watermarking techniques. a watermark should adhere to a few requirements. and source. including intentional attacks to remove the watermark. and transparent. But whereas steganography aims for imperceptibility to human senses. falling into 2 main categories. The media of focus in this paper is the still image. It just marks data.2. A watermark is embedded into a digital signal at each point of distribution. to authenticate the content.Both steganography and digital watermarking employ steganographic techniques to embed data covertly in noisy signals. One application of digital watermarking is source tracking. and common signal processing alterations used to make the data more efficient to store and transmit. depending on in which domain the watermark is constructed: the spatial domain (producing spatial watermarks) and the frequency domain (producing spectral watermarks). the owner can still be identified. then the watermark may be retrieved from the copy and the source of the distribution is known. The signal. However. . Watermarks of varying degrees of obtrusiveness are added to presentation media as a guarantee of authenticity. and makes detection. . by pirates less possible. can be used later to identify the owner of the work. Robustness requires that it be able to survive any alterations or distortions that the watermarked content may undergo. Transparency requires a watermark to be imperceptible so that it does not affect the quality of the content. To be effective in its purpose. but does not degrade it nor controls access to the data. Within this category. The effectiveness of a watermark is improved when the technique exploits known properties of the human visual system. Since a digital copy of data is the same as the original. ownership.1 Principle of digital watermarks A watermark on a bank note has a different transparency than the rest of the note when a light is shined on it. Digital watermarking is the process of inserting a digital signal or pattern (indicative of the owner of the content) into digital content. quality. it should be robust. and therefore removal. These are known as perceptually based watermarking techniques. This technique reportedly has been used to detect the source of illegally copied movies. the class of image-adaptive watermarks proves most effective. This is so that afterwards. known as a watermark. and to trace illegal copies of the work.1. If a copy of the work is found later.

including writers. This suggests that there are many applications that could require image watermarking. An example of an area where copyright protection needs to be enforced is in the on-line music industry. digital libraries. they can be embedded in content with a variety of file formats. surveillance imaging. RIAA reports that "record labels see watermarking as a crucial piece of the copy protection system. and is therefore probable that duplication on the Internet occurs without the rightful owners' permission. The Recording Industry Association of America (RIAA) says that the value of illegal copies of music that are distributed over the Internet could reach $2 billion a year. They are of the opinion that any encryption system can be broken. so that the source as well as the intended recipient of the digital work is known. 2. they all digitally write desired information directly onto images or audio data in such a manner that the images or audio data are not damaged. sooner or later. Copyright owners can incorporate . Furthermore. Digital watermarking is the content protection method for the multimedia era. video-ondemand systems. which has resulted in pirating. Embedding a watermark should not result in a significant increase or reduction in the original data. When digital cameras are used to snapshot an event.3 PURPOSES OF DIGITAL WATERMARKS Watermarks are a way of dealing with the problems mentioned above by providing a number of services:  They aim to mark digital data permanently and unalterably. and satellite-delivered video. medical imaging. Basically.Currently there are various techniques for embedding digital watermarks. Another scenario in which the enforcement of copyright is needed is in newsgathering. However. image and video databases. including Internet imaging. image's origin and content can be verified.1.2 IMPORTANCE OF DIGITAL WATERMARKS The Internet has provided worldwide publishing opportunities to creators of various works. and that digital watermarking is needed to indicate who the culprit is. these same opportunities provide ease of access to these works. 2. the images must be watermarked as they are captured. This is so that later. Digital watermarking is being recognized as a way for improving this situation.1. Digital watermarks are added to images or audio data in such a way that they are invisible or inaudible Ñ unidentifiable by human eye or ear. It is easy to duplicate audio and visual files. whether their music is released over the Internet or on DVD-Audio". digital cameras. musicians and artists. photographers.

translation.identifying information into their work. A/D conversion Resampling. That is. they demonstrate the quality and assure the authenticity of the work.  With a tracking service. The various applications are:    Authentication Broadcast Monitoring Copy Prevention . this section identifies digital watermarking applications and provides an overview of digital watermarking capabilities and useful benefits to customers. In addition. scaling and cropping. watermarks are used in the protection of ownership. Requantization.1. The presence of a watermark in a work suspected of having been copied can prove that it has been copied.5 DIGITAL WATERMARKING APPLICATIONS Digital watermarking is rapid evolving field. D/A conversion. Recompression Linear filtering such as high pass and low pass filtering.  Some more recent techniques are able to correct the alteration as well. Geometric Distortions: include such operations as rotation.  By indicating the owner of the work. Addition of a constant offset to the pixel values Local exchange of pixels other intentional attacks: Printing and Rescanning Watermarking of watermarked image (rewatermarking) 2.4 ATTACKS ON WATERMARKS            Lossy Compression: Many compression schemes like JPEG and MPEG can potentially degrade the data’s quality through irretrievable loss of data. 2. any unauthorized copies that s/he has distributed can be traced back to him/her. owners are able to find illegal copies of their work on the Internet. Common Signal Processing Operations: They include the followings. because each purchaser of the data has a unique watermark embedded in his/her copy.1.  Watermarks can be used to identify any changes that have been made to the watermarked data.

further.2 DIGITAL SIGNATURE A digital signature or digital signature scheme is a mathematical scheme for demonstrating the authenticity of a digital message or document. while also claiming their private key remains secret.[1] but not all electronic signatures use digital signatures. video e. Digital signatures are equivalent to traditional handwritten signatures in many respects. meaning that the signer cannot successfully claim they did not sign a message. copyright protection. and must be implemented properly to be effective. For messages sent through a nonsecure channel. Digital signatures are commonly used for software distribution. a properly implemented digital signature gives the receiver reason to believe the message was sent by the claimed sender. and internet tracking. India. and in other cases where it is important to detect forgery or tampering. financial transactions.[2][3] In some countries.t. Stegnosign: For creating. Digital signatures can also provide nonrepudiation. . but properly implemented digital signatures are more difficult to forge than the handwritten type. Digital signatures employ a type of asymmetric cryptography.6 WATERMARKING SOFTWARE&SREVICES      Alpha-Tec: watermarking software for copyright protection and infringement tracking.  Forensic Tracking E-Commerce/Linking 2. the signature is valid.1. Signum: Allow digital fingerprints to be embedded into grahics. embedded messages and more. A valid digital signature gives a recipient reason to believe that the message was created by a known sender such that they cannot deny sending it (authentication and non-repudiation) and that the message was not altered in transit (integrity). Digital signature schemes in the sense used here are cryptographically based. MediaSec: Provide software for various media types.[4] and members of the European Union. audio. electronic signatures have legal significance.c. a broader term that refers to any electronic data that carries the intent of a signature. so that even if the private key is exposed. 2. some non-repudiation schemes offer a time stamp for the digital signature. Digimarc: For document verification. including the United States. partial encryption. Digital signatures are often used to implement electronic signatures. embedding and detecting watermarks.

For example. The algorithm outputs the private key and a corresponding public key. First. Below are some common reasons for applying a digital signature to communications: 2. and status of an electronic document as well as acknowledging informed consent and approval by a signatory. a signature generated from a fixed message and fixed private key should verify the authenticity of that message by using the corresponding public key.  A signature verifying algorithm that.2. either accepts or rejects the message's claim to authenticity. Digital signatures can be used to authenticate the source of messages. produces a signature. given a message. The importance of high confidence in sender authenticity is especially obvious in a financial context. a valid signature shows that the message was sent by that user.1 Uses of Digital Signature As organizations move away from paper documents with ink signatures or authenticity stamps. Secondly. When ownership of a digital signature secret key is bound to a specific user. Two main properties are required. public and private laws. it should be computationally infeasible to generate a valid signature for a party who does not possess the private key. digital signatures can provide added assurances of the evidence to provenance. identity. suppose a bank's branch office .Digitally signed messages may be anything representable as a bitstring: examples include electronic mail. public key and a signature.1 Authentication Although messages may often include information about the entity sending a message. University of Chicago.  A signing algorithm that. 2. Universities including Penn State. or a message sent via some other cryptographic protocol. and congressional bills with digital signatures. given a message and a private key.2. that information may not be accurate. contracts. A digital signature scheme typically consists of three algorithms:  A key generation algorithm that selects a private key uniformly at random from a set of possible private keys. The United States Government Printing Office (GPO) publishes electronic versions of the budget. and Stanford are publishing electronic student transcripts with digital signatures.1.

there is no efficient way to modify a message and its signature to produce a new message with a valid signature. 2. access to the public key only does not enable a fraudulent party to fake a valid signature. acting on such a request could be a grave mistake.2 Integrity In many scenarios.1. prevent this. Similarly. The device signature may be in the form of  sensor pattern noise (SPN)  camera response function  Re sampling artifacts  Color filter array  Interpolation artifacts  JPEG compression  Lens aberration  sensor dust . or more specifically non-repudiation of origin. If the central office is not convinced that such a message is truly sent from an authorized source. it may be possible to change an encrypted message without understanding it. any change in the message after signature will invalidate the signature. By this property. Furthermore.3 Non-repudiation Non-repudiation.2. the sender and receiver of a message may have a need for confidence that the message has not been altered during transmission. an entity that has signed some information cannot at a later time deny having signed it.1.2. 2.) However. known as nonmalleable ones. (Some encryption algorithms. is an important aspect of digital signatures. but others do not.sends instructions to the central office requesting a change in the balance of an account. because this is still considered to be computationally infeasible by most cryptographic hash functions (see collision resistance). Although encryption hides the contents of a message. if a message is digitally signed.

absolute color space. The responses of the filters do not generally correspond to the CIEcolor matching functions. green. Since sensors are made of semiconductors they obey solid-state physics. For example. The sensor's passbandquantum efficiency and span of the CFA's spectral responses are typically wider than the visible spectrum. such that the separate filtered intensities include information about the color of light. The raw image data captured by the image sensor is then converted to a full-color image (with intensities of all three primary colors represented at each pixel) by a demosaicing algorithm which is tailored for each type of color filter. is a mosaic of tiny color filters placed over the pixel sensors of an image sensor to capture color information. The spectral transmittance of the CFA elements along with the demosaicing algorithm jointly determine the color rendition. In photography. . so a color translation is required to convert the tristimulus values into a common. 1 blue and 1 red filter. Each two-by-two submosaic contains 2 green. Color filters are needed because the typical photosensors detect light intensity with little or no wavelength specificity. a color filter array (CFA). The color filters filter the light by wavelength range. and blue (RGB) wavelength regions. thus all visible colors can be distinguished. or color filter mosaic (CFM).CHAPTER 3 COLOR FILTER ARRAY 3. each covering one pixel sensor. the Bayer filter (shown to the right) gives information about the intensity of light in red. Color filter array The Bayer color filter mosaic. and therefore cannot separate color information.

and one magenta. one red. One cyan. green. used in a few cameras of Kodak. Bayer filter RGBE filter CYYM filter CYGM filter RGBW Bayer RGBW #1 2×2 2×2 2×2 2×2 RGBW #2 RGBW #3 Three example RGBW filters from Kodak. and red sensors on top of each other. This arrangement does not require a demosaicing algorithm because each pixel has information about each color. Dick Merrill of Foveon distinguishes the approaches as "vertical color filter" for the Foveon X3 versus "lateral color filter" for the CFA. two yellow. Traditional RGBW similar to Bayer and RGBE patterns. With one blue.1 Manufacture of the CFA .  List of color filter arrays Pattern (pixels) 2×2 size Image Name Description Very common RGB filter. and one magenta.The Foveon X3 sensor uses a different structure such that a pixel utilizes properties of multijunctions to stack blue. one green. used in a few cameras. with 50% white. one yellow. Bayer-like with one of the green filters modified to "emerald". and two green. used in a few Sony cameras. (See Bayer filter#Alternatives) 4×4 2×4 3. One cyan.

thicknesses ranging up to 1 micrometre are readily available. Ensuring that the dyes have the right mechanical properties—such as ease of application. types. For instance. and optimal manufacturing processes of color filter arrays. at best. This makes it difficult. and subsequently dyed.".. A few sources discuss other specific chemical substances. but nowadays. Theuwissen says "Previously. durability. the color filter was fabricated on a separate glass plate and glued to the CCD (Ishikawa 1981). . 1982). CMCR106R. attending optical properties.Diazonaphthoquinone (DNQ)-novolacphotoresist is one material used as the carrier for making color filters from color dyes. CMCR101B. Color photoresists sometimes used include those with chemical monikers CMCR101R. state "Several factors influence the CFA's design. First. Pigment based CFAs have become the dominant option because they offer higher heat resistance and light resistance compared to dye based CFAs. Nakamura said that materials for on-chip color filter arrays fall into two categories: pigment and dye. There is some interference between the dyes and the ultraviolet light needed to properly expose the polymer. Filter materials are manufacturer specific. and spatial frequencies of the absorptive filters. to fine-tune the spectral responsivities. with the G filter being an overlap of the Y and C filters.[15] The gelatin is sectionalized. CMCR106G. via photolithography. Theuwissen makes no mention of the materials utilized in CFA manufacture. and resistance to humidity and other atmospheric stresses—is a challenging task. Adams et al. Aoki reveals that a CYWG arrangement was used. Some sources indicate that the CFA can be manufactured separately and affixed after the sensor has been manufactured. aliasing effects. and CMCR106B. 1978) and not as a hybrid." He provides a bibliography focusing on the number. while other sensors have the CFA manufactured directly on the surface of the imager. all single-chip color cameras are provided with an imager which has a color filter on-chip processed (Dillon. though solutions have been found for this problem. the individual CFA filters are usually layers of transmissive (absorptive) organic or pigment dyes. CMCR101G. In either case. moire patterns. At least one early example of an on-chip design utilized gelatin filters (Aoki et al.

# 61551. the later stages of the integrated circuit manufacturing line). 3. In U.# 4.I. Carl Chiulli cites the use of 5 chemicals. organics would be preferred over glass. On the other hand.P.Given that the CFAs are deposited on the image sensor surface at the BEOL (back end of line.S.501. Ocean Optics has indicated that their patented dichroic filter CFA process (alternating thin films of ZnS and Cryolite) can be applied to spectroscopic CCDs. some CVD silicon oxide processes are low temperature processes.3 IMAGE NOISE Image noise is random (not present in the object imaged) variation of brightness or color information in images. # 5. or Colour Index numbers. Gersteltec sells photoresists that possesses color filter properties. and C. Image noise is an undesirable by-product of image capture that adds spurious and extraneous information. 3. Image noise can also originate in film grain and in the unavoidable shot noise of an ideal photon detector. Solvent Yellow 88.808. list some 150-200 chemical structures.S. mainly azo dyes and pyrazolone-diazenyl. . Solvent Blue 36. CAS Registry numbers.2 Some pigment and dye molecules used in CFAs In U.801 Koya et al. of Fuji Photo Film company. but fail to provide chemical names. AKA Solvent Red 8..I. and is usually an aspect of electronic noise. where a low-temperature regime must be rigidly observed (due to the low melting temperature of the aluminum metalized "wires" and the substrate mobility of the dopants implanted within the bulk silicon).P. three of which are C. It can be produced by the sensor and circuitry of a scanner or digital camera. #12715.096.

Amplifier noise is a major part of the "read noise" of an image sensor. unwanted electrical fluctuations in signals received by AM radios caused audible acoustic noise ("static"). independent at each pixel and independent of the signal intensity. By analogy unwanted electrical fluctuations themselves came to be known as "noise. inaudible. there can be more noise in the blue channel.2 Salt-and-pepper noise . 3. of the constant noise level in dark areas of the image.4.1 Amplifier noise (Gaussian noise) The standard model of amplifier noise is additive." Image noise is. The magnitude of image noise can range from almost imperceptible specks on a digital photograph taken in good light. that is. 3. of course.4. to optical and radioastronomical images that are almost entirely noise. from which a small amount of information can be derived by sophisticated processing (a noise level that would be totally unacceptable in a photograph since it would be impossible to determine even what the subject was). In color cameras where more amplification is used in the blue color channel than in the green or red channel. including that which comes from the reset noise of capacitors ("kTC noise").Noise clearly visible in an image from a digital camera The original meaning of "noise" was and remains "unwanted signal". Gaussian. caused primarily by Johnson–Nyquist noise (thermal noise).4 Types o o o o o o o Amplifier noise (Gaussian noise) Salt-and-pepper noise Shot noise Dark current noise Quantization noise (uniform noise) Read noise Anisotropic noise 3.

display. The variable dark charge of normal and hot pixels can be subtracted off (using "dark frame subtraction"). and the noises at different pixels are independent of one another. If dark-frame subtraction is not done. Shot noise follows a Poisson distribution. or random component. This noise is known as photon shot noise. etc. there can be additional shot noise from the dark leakage current in the image sensor. It can be mostly eliminated by using dark frame subtraction and interpolating around dark/bright pixels. 3.3 Shot noise The dominant noise in the lighter parts of an image from an image sensor is typically that caused by statistical quantum fluctuations. In addition to photon shot noise.Image with salt and pepper noise Fat-tail distributed or "impulsive" noise is sometimes called salt-and-pepper noise or spike noise. Dark current is greatest at "hot pixels" within the image sensor. Shot noise has a root-mean-square value proportional to the square root of the image intensity. that is. bit errors in transmission. or if the exposure time is long enough that the hot pixel charge exceeds the linear charge capacity. This type of noise can be caused by analog-to-digital converter errors. variation in the number of photons sensed at a given exposure level. this noise is sometimes known as "dark shot noise" or "dark-current shot noise". which is usually not very different from Gaussian. Dead pixels in an LCD monitor produce a similar. leaving only the shot noise. but non-random. of the leakage. and hot pixels appear as salt-and-pepper noise. the noise will be more than just shot noise.4. . An image containing salt-and-pepper noise will have dark pixels in bright regions and bright pixels in dark regions.

5 In digital cameras . 3. Calibrating the read noise allows us know more about the quality of the CCD as well as the data distortion due to the reading of images. since high read noise decreases the quality of the image. The read noise of a camera affects how well the image represents the actual data. with states near mid-band generating most of the dark current.[13] 3. Though it can be signal dependent. the most effective way to reduce dark current is to cool the CCD. For example. These sites introduce electronic states in the forbidden gap which act as steps between the valence and conduction bands.4. adding to the signal measured in the pixel. from which they are emitted into the conduction band. The efficiency of a generation center depends on its energy level.5 Quantization noise (uniform noise) The noise caused by quantizing the pixels of a sensed image to a number of discrete levels is known as quantization noise. 3. It has an approximately uniform distribution. image sensors are sometimes subject to row noise or column noise. it will be signal independent if other noise sources are big enough to cause dithering. robbing electrons of the thermal energy required to reach an intermediate state.3. providing a path for valence electrons to sneak into the conduction band. For this reason. or if dithering is explicitly applied.4.4.4.6 Read noise Read noise is a property that is inherent to the CCD of digital cameras.4 Dark current noise: Dark current is the result of imperfections or impurities in the depleted bulk silicon or at the siliconsilicon dioxide interface. 3. The generation of dark current is a thermal process wherein electrons use thermal energy to hop to an intermediate state. and is present in all images taken and recorded by a camera.7 Anisotropic noise Some noise sources show up with a significant orientation in images.

The relative effect of both read noise and shot noise increase as the exposure is reduced. At the cost of a doubling of read noise variance (41% increase in read noise standard deviation). longer shutter speeds lead to increased salt-and-pepper noise due to photodiodeleakage currents. or both. or effective light collection area per pixel sensor. the sensitivity of an imager scales roughly with the sensor area. In low light. Banding noise. so larger sensors typically create lower noise images than smaller sensors. or printed at the same size. this salt-and-pepper noise can be mostly eliminated by dark frame subtraction. In the case of images bright enough to be in the shot noise limited regime. the pixel count makes little difference to perceptible noise levels – the noise depends primarily on sensor area. can be introduced through brightening shadows or through color-balance processing. or that the f-number or focal-plane illuminance is held constant. assuming the aperture area is proportional to sensor area. On most cameras. where read noise (noise floor) is significant. correct exposure requires the use of long shutter speeds. For images at lower signal levels (higher ISO settings). since fewer photons are counted (shot noise) and since more amplification of the signal is necessary. more pixels within a given sensor area will make the image noisier if the per pixel read noise is the same. 3. for a constant f-number. when the image is scaled to the same size on screen.Image on the left has exposure time of >10 seconds in low light. That is. higher gain (ISO sensitivity). The image on the right has adequate lighting and 0. similar to shadow noise.1 second exposure. . corresponding to increased ISO sensitivity. not how this area is divided into pixels.6 Effects of sensor size The size of the image sensor. is the largest determinant of signal levels that determine signal-to-noise ratio and hence apparent noise levels.

it may be more likely to be a true feature of the image.5" compact camera sensor (with roughly 1/16 the area) at ISO 100. However.7 Sensor heat Temperature can also have an effect on the amount of noise produced by an image sensor due to leakage. Many cameras have settings to control the aggressiveness of the in-camera noise reduction. many people find . but not strictly noise). involve some form of noise reduction. Most noise reduction algorithms perform much more aggressive chroma noise reduction. it is known that DSLRs will produce more noise during summer than winter. and average out the former while attempting to preserve the latter.8 Image noise reduction Most algorithms for converting image sensor data to an image. if it covers a few pixels in an absolutely regular shape. A simplified example of the impossibility of unambiguous noise reduction: an area of uniform red in an image might have a very small black part. no algorithm can make this judgment perfectly. This ability to produce acceptable images at higher sensitivities is a major factor driving the adoption of DSLR cameras. whether in-camera or on a computer. With this in mind. since there is little important fine chroma detail that one risks losing. so there is often a tradeoff made between noise removal and preservation of fine. low-contrast detail that may have characteristics similar to noise.For instance. This decision can be assisted by knowing the characteristics of the source image and of human vision. 3. But a definitive answer is not available. and that produced by a 1/2. An example shows a DSLR sensor at ISO 400 creating less noise than a point-and-shoot sensor at ISO 100. which tend to use larger sensors than compacts. it is likely (but not certain) to be spurious and noise. if it is irregular. Furthermore. it may be a defect in a group of pixels in the image-taking sensor (spurious and unwanted. 3. If this is a single pixel. but all attempt to determine whether the actual differences in pixel values constitute noise or real photographic detail. There are many procedures for this. the noise level produced by a Four Thirds sensor at ISO 800 is roughly equivalent to that produced by a full frame sensor (with roughly four times the area) at ISO 3200.

. since its textured appearance mimics the appearance of film grain. it is not so much a matter of noise reduction as of extracting a little information buried in a lot of noise. such as astronomical images of very distant objects. In cases of extreme noise.luminance noise less objectionable to the eye. Since noise levels increase as ISO sensitivity is increased. The high sensitivity image quality of a given camera (or RAW development workflow) may depend greatly on the quality of the algorithm used for noise reduction. most camera manufacturers increase the noise reduction aggressiveness automatically at higher sensitivities. techniques are different. seeking small regularities in massively random data. This leads to a breakdown of image quality at higher sensitivities in two ways: noise levels increase and fine detail is smoothed out by the more aggressive noise reduction.

which might strongly deteriorate if additional electrical gain (and noise) is included. it may vary with integration time. The term "fixed pattern noise" usually refers to two parameters.[1] One is the DSNU (dark signal non-uniformity). degrading the image. which is the offset from the average across the imaging array at a particular setting (temperature. material or interference with the local circuitry. it is not expressed in a random (uncorrelated or changing) spatial distribution. In practice.CHAPTER 4 FIXED PATTERN NOISE Fixed pattern noise is the term given to a particular noise pattern on digital imaging sensors often noticeable during longer exposure shots where particular pixels are susceptible to giving brighter intensities above the general background noise. integration time) but no external illumination and the PRNU (photo response nonuniformity). pixel dependent photo response nonlinearity (PRNL) and is often simplified as a single value measured at almost saturation level to permit a linear approximation of the non-linear pixel response. It is characterized by the same pattern of 'hot' (brighter) and cold (darker) pixels occurring with images taken under the same illumination conditions in an imaging array. Sometimes pixel noise[2] as the average deviation from the array average under different illumination and temperature conditions is specified. which describes the gain or ratio between optical power on a pixel versus the electrical signal output. The latter can be described as the local. This problem arises from small differences in the individual responsibility of the sensor array (including any local post amplification stages) that might be caused by variations in the pixel size. occurring only at certain. Pixel noise therefore gives a number (commonly expressed in rms) that identifies FPN in all permitted imaging conditions. imager temperature. Although FPN does not change appreciably across a series of captures. and fixed pixel locations. It might be affected by changes in the environment like different temperatures. imager gain and incident illumination. a long exposure (integration time) emphasizes the inherent differences in pixel response so they may become a visible defect. etc. exposure times. Fixed pattern noise (FPN) is a general term that identifies a temporally constant lateral nonuniformity (forming a constant pattern) in an imaging system with multiple detector or picture elements (pixels). .

00 candelas since the brighter light is more easily detected by the camera. One of the few engineering definitions for PRNU or "photoresponsenonuniformity" is in the photonics dictionary. The calculation of the PRNU is as follows:  Obtain the average image over the 100 images taken:   Subtract the DSNU image from this average image to eliminate the contribution from the DSNU. and increasing the illumination level increases the non-uniformity of the illumination .5. is one source of pattern noise in digital cameras. Repeat the calculations for the different exposure times to compare the PRNU.1.2 Methods To characterize the PRNU. We took 100 exposures each for three exposure times: 1/10. Obtain the spatial variance of the pixel values over the entire CCD   Divide the spatial variance by the average image from (ii) to obtain the PRNU as a percentage of the actual pixel values. 4. PRNU is the pixel variation under illumination. produced by the Optoliner. 4. We kept the illumination level fixed at 3. Like DSNU. since increasing the illumination level will enhance the difference in the photo-response of the pixels across the image and lead to a higher PRNU.1 Background Photo Response Non-Uniformity. or PRNU for short. However. since the maximum value of the Opt linear device is around 4 candelas. In our measurements.1. it is seen as the variation in pixel responsively over the CCD. We expect the PRNU to increase with increasing illumination. And it is for CCD only.1 PRNU (Photo Response Non-Uniformity) 4. 1/4 and 1/2. and we also checked to ensure that the camera is focused before taking the pictures. we use the camera to take multiple images of a uniform scene. while DSNU occurs as a variation in pixel responsively when the CCD is not illuminated.

a camera output model can be expressed as where I is the output image. read-out noise and quantization (lossy Compression) noise. G. (1) and keeping the first two terms of the expansion where  is the denoised image and is the ensemble of the noises. The PRNU pattern noise K can then be formulated as . including . The dominating component of sensor pattern noise is photo response non-uniformity (PRNU). (1). However.e. A colour photo is represented in three colour components (i. shot noise.produced by the Opt linear. the signal will inevitably be distorted when passing through each process and these distortions result in slight differences between the scene and the camera-captured image. stand for dark current. K is the zero-mean multiplicative factor responsible for the PRNU. . (= 0. we chose to increase the exposure times to mimic the effect of increasing illumination levels. respectively.. a colour interpolation function generates the electronic signals of the other two colour components for every pixel according to the colour intensities of the neighboring pixels. Since is the dominating term in Eq. the PRNU can be contaminated by various types of noise introduced at different stages of the image acquisition process. (1). As formulated in [11]. the lenses let through the rays of the three colour components of the scene. during the image acquisition process. but for every pixel only therays of one colour component is passed through the CFA and subsequently converted into electronic signals by the sensor. This colour filtering is determined by the CFA. This colour interpolation process is commonly known as demosaicking. In acquiring an image. Finally.455) is the gamma correction factor. R. The signals then undergo additional signal processing such as white balance. In Eq. after applying Taylor expansion to Eq. After the conversion. these signals are stored in the camera’s memory in a customized format.s andr are random noise and is the fixed pattern noise (FPN) that is associated with every camera and can be removed by subtracting a dark frame from the image taken by the same camera. and is the input signal of the scene. and B). gamma correction and image enhancement. and . g is the colour channel gain. Figure 1 demonstrates the image acquisition process. primarily the JPEG format. For most digital cameras.

(6) is element-wise. (5) is element-wise.2 Use of PRNU in Device Identification The basic idea of using the PRNU noise pattern in device identification can be described as follows. it is important that the PRNU extracted is as close to the genuine pattern noise due to the sensor as possible. thus identifying the source device that has taken the image under investigation. the noise residual patterns are extracted using Eq. i. The normalised cross-correlation is the noise residual extracted from . is used to compare the noise Note in Eq. 1) First. the discrete wavelet transform followed by a Wiener filtering operation). the wavelet-based denoising process (i. has been reported as effective in producing good results. where S is the number of images involved in the calculation. (5) and compared against the reference PRNU Kd of each device d available to the investigator in the hope that it will match one of the reference fingerprints. instead of using against the reference fingerprint .is the noise residual obtained by applying a denoising filter on image I. 4. Given the PRNU-based approaches‟ potential in resolving device identification problem to the accuracy at individual device level. we used . only one of the three colours of each pixel is physically captured by the sensor while the other two are artificially interpolated by the demosaicking process. Again the multiplication operation in Eq.e. (6). Since for most cameras.. is the gamma correction factor .is the s-th image taken by device d and Note the multiplication operation in Eq.e. Although various denoising filters can be used.. (5) from a number of low-contrast images taken by device d and then the PRNU is estimated using the ML estimation procedure adopted by Chen et. as suggested in [11]. 2) Secondly. the noise residual WI of image I under investigation is extracted using Eq. al. this inevitably introduce noise with power stronger than that of the genuine .. where is the mean function. for each imaging device d.

Addressing this shortcoming is the motivation of this work. interchangeably whenever there is no need to differentiate them. such as chromatic aliases.PRNU. In this work. Because the PRNU is formulated in Eq. In Section IV. The algorithm should have the following traits:  Avoidance of the introduction of false color artifacts. Most modern digital cameras acquire images using a single image sensor overlaid with a CFA. 4. rather than using the camera's built-in firmware. PRNU and noise residual. Eq. we will look at the impact of demosaicking on PRNU fidelity in Section II and propose an improved formula for extracting PRNU in Section III. We can see from Eq.e.3 DEMOSAICING A demosaicing (also de-mosaicing or demosaicking) algorithm is a digital image process used to reconstruct a full color image from the incomplete color samples output from an image sensor overlaid with a color filter array (CFA). Section V concludes this work.e. (4)). However.. Many modern digital cameras can save images in a raw format allowing the user to demosaic it using software. we present some experiments on device identification and image content integrity verification to validate the proposed PRNU extractionformula. (3) and (5) as a function of the noise residual W (i. It is also known as CFA interpolation or color reconstruction. allowing the interpolation noise in the artificial components to contaminate the real PRNU in the physical components. a full set of color triples) from the spatially under sampled color channels output from the CFA. The aim of a demosaicing algorithm is to reconstruct a full color image (i. as mentioned earlier that I is to apply the discrete wavelet transform followed by a Wiener filtering operation directly to the entire image I without differentiating physical components from artificial components and. (3) and (4) that the accuracy of both PRNU K and noise residual W depends on the denoising operation applied to I in obtaining the most common method of obtaining . zippering (abrupt unnatural changes of intensity over a number of neighboring pixels) and purple fringing  Maximum preservation of the image resolution . as a result. so demosaicing is part of the processing pipeline required to render these images into a viewable format. in the rest of the work we will use the two terms. (2).

To extract the PRNU using Eq. When wavelet transform is applied during the noise residual extraction process the interpolation noise residing in the artificial components propagates into the physical components. 4. the discrete wavelet transform followed by a Wiener filtering operation is applied. a form of interpolation is needed to fill in the blanks. (4) and (5) based on the definition of the output image model in Eq. The grouping process is usually content-dependent. (4) is that it involves the whole image plane. (4) and (5). in the red colour component/plane of an image taken by a camera with a Bayer CFA.g. Due to the fact that demosaicking is a key deterministic process that affects the quality of colour images taken by many digital devices. demosaicking has been rigorously investigated. The mathematics here is subject to individual implementation. For example. For example. edge-adaptive or non-adaptive. However. hence the accuracy of colour interpolation result is also content-dependent.4 DEMOSAICKING IMPACT ON PRNU FIDELITY In this work. thus the interpolation noise is also more significant. Conversely. This indicates that the PRNU in physical colour components is more reliable than that in the artificial components. As a result the interpolation noise gets diffused from the artificial components into the physical ones. in one noise residual extraction process. the existing method for extracting PRNU as formulated in Eq. and is called demosaicing. which contains both artificial and physical components. Most demosaicking approaches group the missing colours before applying an interpolation function. we call the colour components physically captured by the sensor as physical colours and the ones artificially interpolated by the demosaicking function as artificial colours. Low computational complexity for fast processing or efficient incamera hardware implementation  Amenability to analysis for accurate noise reduction To reconstruct a full color image from the data collected by the color filtering array. (1) does not take this into account. The main problem inherent to Eq.. e. the colour variation between neighbouring pixels is greater. each coefficient of the wavelet transform used in the noise residual extraction process involves multiple pixels and thus both artificial and physical components. only one fourth of the pixels‟ red colour are physical and for each pixel with physical red colour all its 8 neighbours‟ red colours are artificial. in inhomogeneous areas. the interpolation function can more accurately generate artificial components. because of the low variation of the colour intensities of neighbouring pixels. However. in a homogeneous area. Therefore it is desirable to devise a noise residual extraction method that can .

.prevent the artificial components from contaminating the reliable PRNU residing in the physical components with the interpolation noise.

According to Eq. 5. the second part of Eq. a mathematical model for the CD-PRNUis derived and then an extraction algorithm is proposed to extract the noise residual that is to be used for estimating the final CD-PRNU. (7) means that if the colour component c is the same as the colour that the CFA pattern F allows to pass. For each colour component of a pixel . the image output model of Eq. only one of the three colour components takes a value physically captured by the sensor and this colour is determined by the colour configuration of the CFA pattern F.1 FORMULATION OF COLOUR DECOUPLED PRNU (CD-PRNU) In this section. The other two colour components are to be determined by the demosaicking process. we will discuss the formulation and extraction of CD-PRNU.e . (1) proposed in can be reformulated as . i. Otherwise. F. Let be an interpolation matrix with 2N+1 × 2N+1 coefficients and be a X × Y-pixel input signal from the scene consisting of three colour components. (7) is artificially applied to calculate the colour. (7). can be determined according to The first part of Eq. G (green) and B (blue) before colour interpolation. then no demosaicking is needed because c has been physically captured by the sensor. we makes no assumption about the CFA pattern. R (red).2 Mathematical Model of CD-PRNU A generic demosaicking process is to convolve an interpolation matrix with an image block of the same size centred at the pixel where the artificial colour is to be calculated. to make the proposed CD-PRNU versatile and applicable to cameras adopting different CFA patterns. First. except that it is a 2 × 2 square array. Although the 2×2 Bayer CFA is the most common CFA pattern. That is to say that for each pixel .CHAPTER 5 CD-PRNU (Color Decoupled Photo Response Non-Uniformity) 5. without prior knowledge about the CFA.

Hence if the physical and artificial colour components can be separated / decoupled. 5. We can also use the same ML estimate as in Eq.. Eq. (5) to extract the reference interpolation noise d from S low-variation images taken by d such that for a particular device where is the artificial colour components of the s-th low-contrast image taken by device d and . in the physicalcomponents. (9) suggests that in the artificial components. (3)). the PRNU is actually the interpolation noise P while. which is actually the interpolation noise. (9) that the physical components and artificial components have similar mathematical expression.3 CD-PRNU Extraction Algorithm . That is where is a low-passed filtered version of the artificial components and is the corresponding ―sensor pattern noise‖. the PRNU remains unaffected by the interpolation noise.Eq. P can be extracted in the same way as the sensor pattern noise K is extracted (i.e. We will discuss how the physical and artificial is the interpolation noise extracted from colour components can be decoupled in simple manner without a priori knowledge about the CFA pattern in Section III. It can also be seen from Eq.B.

Unfortunately. . we do not know (actually we do not have to know) which pixels carry the colour captured physically by the hardware and which are not. In this work. respectively. these methods have to exhaust all of the possible CFA patterns in order to infer/estimate the „real‟/optimal CFA. By decoupling the physical and artificial colour components in this fashion before extracting the noise residual. The framework of the colour decoupled noise residual extraction process is shown in Figure 2 and the procedures are listed in Algorithm 1. to extract the CD-PRNU. manufacturers usually do not provide information about the CFA used by their cameras. several methods have been proposed to estimate the CFA. . and mod is the modulo operation. However. from the physical and artificial components if the CFA is known. we first separate the three colour channels of a colour image I of pixels. For each colour channel. We know that. we perform a 2:1 down-sampling across both horizontal and . (11) is applied. Most CFA patterns are of 2 × 2 elements and are periodically mapped to the sensors. Eq. we can prevent the artificial components from contaminating the physical components during the DWT process. . (10) and (11). so the second step is. Note that Algorithm 1 is for extracting the noise residual pattern W from an image I. only one of the three colour components is physical and the other two are artificial. Finally the CD-PRNU Wc of each colour channel c is formed by combining the four sub-noise residuals such that where. Eq. without knowing the CFA pattern used by the manufacturer.According to Eq. However. But by decomposing into four sub-images. such that vertical dimensions to get four sub-images. exhaustive search is by no means acceptable. we know that each of the four sub-images either contains only the physical colour or only the artificial colours. for each pixel of I. we can extract the sensor pattern noise and interpolation noise. for each channel . To estimate the CD-PRNU Pd of a particular device d and use it as the reference signature of d. (4) is then used to obtain noise residual from each sub-images . Therefore.

.

5 EXPERIMENTAL RESULTS In this section. we need clean PRNUs (which appear as high frequency bands of images) as device . The six cameras are listed in Table1. For device identification purpose. Source Camera Identification We have carried out source camera identification tests on 300 2048×1536-pixel photos of natural scenes taken by six cameras(C1 to C6). Noise residual extraction algorithm Input: original image I Output: colour decoupled noise residual W Noise residual extraction algorithm 5. Table 1. each responsible for 50.4 Algorithm 1. ) of each camera Ci is generated by taking the weighted average of the PRNUs extracted from 30 photos of blue sky according to Eq. Cameras used in the experiments.1. 5.5. The reference PRNU (i. we carry out experiments on source camera identification and image content integrity verification to validate the feasibility of the proposed CD-PRNU in a comparative manner. (11).5.e.

For C2. (6) will be used to measure the similarity between PRNUs (CD-PRNUs).fingerprints for comparison against the PRNU extracted from individual images under investigation. In practice. other images with low-variation scenes (i. C3. Table 2 lists the identification rates. so geometrical transformations will not be applied in order to prevent biased evaluation from happening. Generally speaking. Therefore. Fridrich suggested the use of the Peak to Correlation Energy (PCE) measure in [15]. thus giving better chance of extracting clean PRNU. Therefore. the normalised cross-correlation has to be greater than a specified threshold for a camera to be identified as the source camera. scenes without significant details) can be used instead. PRNU performs equally well as CD-PRNU when the image size is 192 × 256 pixels and slightly outperforms CD-PRNU when the block size is 48 × 64 pixels. its reference PRNU (or CD-PRNU) is most similar to the PRNU (or CD-PRNU). C5 and C6 perform significantly better when CD-PRNU is used in all cases. the less data is available. normalised cross-correlation formulated as in Eq. However.. a camera is identified as the source camera. except for a few cases when images are of full size (1536 × 2048 pixels) and the identification rates are close or equal to 100% (1. C1.0000). therefore identification results become less reliable. the key point is about demonstrating the different performance of the traditional PRNU and the proposed CD-PRNU. C4. such as rotation or scaling. Another reason is that. in this experiment. Our empirical experience suggests that an average of 20 blue sky images is accurate enough. Taking the average of the PRNUs from 30 blue sky images is to further reduce variation. Actually. in the following experiments. if out of the six reference PRNUs (or CDPRNUs). Individually speaking.e. of the image I under investigation. The purpose of this experiment is to demonstrate the capability of the proposed CD-PRNU in dealing with the colour interpolation noise. We suspect that the reason C2 does not perform as expected is because the CFA pattern is not a 2 × 2 square array as we have assumed. Because PRNU is often used in content integrity verification. because the smaller the images. The reason blue-sky images are chosen in this work is because blue sky contains less scene details (high frequency signal). where smaller image blocks have to be analysed. WI. which has been proved to be a more stable detection statistics than normalised cross-correlation when applied to the scenarios in which the images of interest may have undergone geometrical manipulations. Source camera identification requires similarity comparisons among PRNUs (CD-PRNUs) and therefore the feasibility of the chosen similarity metrics is important. . we also compare the performance of the proposed CD-PRNU against that of the traditional PRNU [11] when they are applied to blocks of 5 different sizes cropped from the centre of the fullsized PRNU (CD-PRNU).

Performance comparison of source camera identification a) Overall identification rates when CD-PRNU and PRNU are used as fingerprint In Figure 3(b). we can see that CD-PRNU still outperforms PRNU significantly. as listed in the Total column of Table 2. . This has been graphically presented in Figure 3(a).when the statistics of the six cameras are pooled together. a ROC curve of the performance of PRNU and CD-PRNU are demonstrated. Figure 3. We can see that the CD-PRNU outperforms the PRNU because at all fixed False Positive rate the CDPRNU‟s True Positive rate are always higher than that of the PRNU.

Figure 3. The amount of data processed during the extraction of PRNU and CD-PRNU is the same. it takes 0. these two operations are trivial and only incur negligible increase of time complexity. Performance comparison of source camera identification b) Overall ROC curve when CDPRNU and PRNU are used as fingerprint For a system with a Pentium Core II 1.567 seconds to calculate the similarity between a pair of CD-PRNUs of the same size.3G CPU and 3 GB RAM. . Although extracting CD-PRNU requires down-sampling and up-sampling.526 seconds to compute the similarity between the PRNUs of two images of 2048 × 1536 pixels and 0.

Table 2. we copied a 160 × 390-pixel area from Image I. .  In the first experiment.5. and pasted it at approximately the same location in Image I. The images in Figure 4(a) and (b) are taken by Olympus C730.2 Content Integrity Verification We also carried out the following three content integrity verification experiments on 640 × 480-pixel images. 5.1 in Figure 4(a).2 in Figure 4(b) to create the forged Image I. Source camera identification rates using traditional PRNU and proposed CD-PRNU.3 as shown in Figure 4(c).

1 in Figure 5(a).1 (b) Original Image II.2. Figure 5.3  In the third experiment. source image and forged images for the content verification experiments.2 (c) Forged Image III. the windows are . The images in Figure 5(a) and (b) are also taken by the same camera.2 (c) Forged Image I. (a) Original Image II.2 in Figure 5(b) to create the forged Image II.3  In the second experiment.1 (b) Original Image I.2 (c) Forged Image II. source image and forged images for the content verification experiments.2 in Figure 6(b). which covers the face of the person.2. which is taken by Olympus C730. we cropped a 60 × 80-pixel area from Image III. The original image. In Chen’s method [11]. Figure 6.1 (b) Original Image III. we cropped an 80 × 100-pixel area from Image II.3 To detect the manipulated areas. II. The original image. we slid a 128 × 128-pixel window across the PRNU extracted from the image under investigation and another window of the same size across the reference PRNU of the cameras that have taken images I. pasted it at the area where the face of another person is in Image III.3 in Figure 5(c). which covers the face of the person.1 in Figure 6(a) taken by Canon Power Shot A400.2 and III.Figure 4. source image and forged images for the content verification experiments. (a) Original Image III. The original image. (a) Original Image I.3 in Figure 6(c). pasted it at the area where the face of another person is in Image II. to create the forged Image III.

. In the following experiments we will allow t to vary independently in the range from 0. therefore. To decide whether a block centered at the window superposed on the image has been manipulated or not. the block in the centre of the window is deemed as manipulated.0 to 3. where and are the mean and standard deviation of the correlations distribution. the cross-correlation of the PRNU patterns inside the two windows at the same location was calculated according to Eq. and lower FP and FN indicate better performance. we will obtain lower TP and FP. TN and FN are defined as TP = |{B | M(B) = 1 and Md(B) = 1}|. and T(t) is the threshold. As discussed in [11]. true negative (TN) and false negative (FN) to measure the performance of integrity verifications based on PRNU and CD-PRNU. true positive (TP).moved a pixel at a time. false positive (FP). while higher TN and FN. Therefore. Let B be an arbitrary block and M(B) and Md(B) be defined as TP. FP = |{B | M(B) = 0 and Md(B) = 1}| and FN = |{B | M(B) = 1 and Md(B) = 0}|. the sliding step/displacement is set to 5 pixels in order to reduce the computational load without sacrificing the accuracy of the integrity verification. we can evaluate the integrity verification performance across a wide range of correlation thresholds T(t). FP. this method is not accurate at the pixel level [11].0 and use the four metrics. which incurs a high computational load. we use various thresholds defined as to analyze the performance of PRNU and CD-PRNU. Table 3. the cross-follows the Generalized Gaussian (GG) distribution. If the cross-correlation is lower than a predetermined threshold t. Higher TP and TN. Table 3 lists the number of manipulated and non-manipulated blocks of 5 × 5 pixels in the forged images. Moreover. in our experiment. respectively. By varying the value of t. TN = |{B | M(B) = 0 and Md(B) = 0}|. As t grows. (6). Number of manipulated and non-manipulated areas in each image (unit: block).

However. FP and FN when authentication is carried out on image I. In order to simplify the comparison and to obtain a fair result. a low threshold also results in the situation where more authentic blocks are mistakenly detected as manipulated.3 across a range of correlation threshold T(t). giving rise to a higher FP. which is marked vertically. which means the authentication result is generally as unreliable as that of a random guess. the experiments on III.2.1 is at approximately the same location as the original area in image I. It is clear that the ROC curve of the PRNU-based scheme mostly overlaps with that of Random Guess. By .3 demonstrate that CD-PRNU-based method significantly outperforms the PRNU-based method when the tampered area is about one quarter of the sliding window. an algorithm with better performance will have a higher true positive rate (). The ROC curves for the integrity verification experiments on image I. TN.3 is illustrated as Figure 8. As a result.According to Chen‟s predication. Therefore a ROCcurve of TP rate with respect to FP rate can be used to evaluate the overall performance of the PRNU and CD-PRNU.2. the performance of the filtering / dilation operation strongly depends on parameter setting and hence many experiments must be run to obtain the best parameters for filtering. FP and FN. the ROC is formulated as At the same false positive rate . TN. therefore the PRNU pattern noises in the two areas are almost the same.1 Experiment on Image I.3 Figure 7 shows the performance of the PRNU and CD-PRNU in terms of TP. 5. which is marked along the horizontal axis of the ROC curve. This is because the area we copied from the source image I. ―the block dimensions impose a lower bound on the size of tampered regions that our algorithm can identify. Thus. As a result. Let α be the number of manipulated blocks and β be the number of authentic blocks. However. we use the raw data without any filtering to calculate the TP. Chen applies erosion and dilation operations with a square kernel in order to filter small areas identified as tampered with. we remove all simply connected tampered regions from Z that contain less than 64×64 pixels (one quarter of the number of pixels in the block)‖.5. the scheme cannot detect the manipulated area based on PRNU. leading to higher TP. We can see from Figure 7(a) and 7(b) that CD-PRNU generally achieves higher TP and TN while maintaining lower FP and FN. A lower correlation (similarity) allows the algorithm to detect more manipulated blocks. The final authentication result is a image with the dilated areas highlighted as the tampered areas.

0. which means that by using CD-PRNU manipulated blocks can be detected more reliably.3 .2.3: Integrity verification performance of the PRNU and CD-PRNU in terms of a) TP. with t varying from 0. the CD-PRNU-based scheme results in a curve much higher than the PRNU-based method.3. across a range of correlation threshold T(t). Figure 7.0 to 3.5.contrast. Figure 8. The ROC curve of Truth Positive Rate with respect to False Positive Rate of PRNU and CD-PRNU when authentication is performed on image I. Authentication results on image I. b) TN. 5.2 Experiment on Image II.

3.0 to 3. . again indicate its superiority to PRNU. as shown in Figure 9(a) and 9(d). both PRNU and CD-PRNU can effectively detect the manipulated blocks as their ROC curves have suggested in Figure 10. Figure 10 also shows that the ROC curve of CD-PRNU is still slightly higher than that of PRNU. Figure 9. mixed performance in terms of TN and FP can be seen in Figure 9(b) and 9(c). Albeit their mixed performance in terms of TN and FP. However. b) TN. CD-PRNU‟s consistently higher TP and lower FN. Authentication results on image II.0. c) FP and d) FN across a range of correlation threshold T(t).When verifying the integrity of image II.3: Integrity verification performance of the PRNU and CD-PRNU in terms of a) TP. indicating a slightly better performance of CD-PRNU. with t varying from 0.

as demonstrated in Figure 11(a). The experiment result on III. As a result. .5. the PRNU cannot correctly detect any manipulated blocks (i. as can be seen in Figure 11(b) and 11(c). the method based on PRNU can perform no better than a random guess. and therefore areas smaller than this should be filtered in order to remove the falsely identified noise.3 When authenticating III.1. Since the tampered area is 60 × 80 pixels. Chen predicated in that one quarter of the sliding window is the lower bound on the size of tampered regions that our algorithm can identify.3 conforms to Chen’s observation. CD-PRNU‟s significantly better performance in terms of TP and lower FN can still be seen again in Figure 11(a) and 11(d). although the performance of PRNU and CD-PRNU in terms of TN and FP are mixed. When the threshold t is higher than 1.2. the CD-PRNU-based method can identify smaller areas. respectively. the manipulated blocks can be effectively detected by the CD-PRNU-based scheme because the areas in question are from two images taken by different cameras and thus contain different interpolation noise. approximately one quarter of the window.Figure 10. By contrast.3. 5. This poor performance is also reflected in the PRNU’s ROC curve in Figure 12 and is due to the fact that he manipulated area is too small (60 × 80 pixels).3 Experiment on Image III.3. which is only about one quarter of the sliding window (128 × 128 pixels). The ROC curve of Truth Positive Rate with respect to False Positive Rate of PRNU and CD-PRNU when authentication is performed on image II.e.

with t varying from 0.Figure 11. b) TN.0 to 3.0. . Authentication results on image III.3: Integrity verification performance of the PRNU and CD-PRNU in terms of a) TP. c) FP and d) FN across a range of correlation threshold T(t).

Master your semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master your semester with Scribd & The New York Times

Cancel anytime.