You are on page 1of 12

DIGITAL SIGNAL PROCESSING LAB

COMPLEX ENGINEERING ACTIVITY

IMAGE DENOISING

Group Members
Humna Sharif (180630)

Ahmad Hassan Saeed (180622)

Khizar Khalid (180606)

Bachelors of Computer Engineering

Fall-2018

Submitted To:

Engr. Sharoze Sohail


Lab Engineer

DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING

FACULTY OF ENGINEERING
Introduction :

In this CEA, an image is provided as a signal which contains noise. The task is to remove the
noise using spatial as well as frequency domain techniques. Noise removal is an important
technique in image processing in which unnecessary data is removed from image using low pass
filter and image smoothing operations. In the spatial domain, noise removal is based on direct
manipulation of pixel values whereas in the frequency domain, the modifying techniques are
based on Fourier transform of images.

Fourier transform:
Discrete Fourier transform is a technique that allows us to convert the signal from spatial to
frequency domain. The reason behind this approach is that in frequency domain, filtering process
is a lot faster as we perform multiplication instead of convolution which is a quite intensive task.
We used the Fast Fourier transform which is the optimized version of the DFT to implement this
problem. The reason behind FFT approach is that DFT is not easy to compute and takes a lot of
time.

Inverse Fourier transform:


Now after filtering and processing of the image in Fourier domain. We need to convert it back to
Spatial domain. Now, this is where inverse Fourier transform comes. Inverse Fourier transform is
a technique that allows us to convert frequency response of a system back into Spatial domain.
We used the same fast Fourier version to implement this problem.

Matlab :

Read Image :

Image can be read by “imread( )” command. Brackets will contain the input image along the
extension.
We have read an image as follows:
Now different filters are performed on the image to reduce noise. We have
performed these following filters.

1. Median Filter :

Median filter is one of the well-known order-statistic filters due to its good performance for
some specific noise types such as “Gaussian,” “random,” and “salt and pepper” noises .
According to the median filter, the center pixel of a M × M neighborhood is replaced by the
median value of the corresponding window. Note that noise pixels are considered to be very
different from the median.
2. Average filter:

Average filtering is a method of ‘smoothing’ images by reducing the amount of intensity


variation between neighbouring pixels. The average filter works by moving through the
image pixel by pixel, replacing each value with the average value of neighbouring pixels,
including itself.
We have tested “Salt & Pepper” noise removal on our image.

Read image :
RGB To Gray :
Salt & Pepper Noise:

Fourier Transform :
Inverse Fourier Transform :
Now applying filter to remove noise,

Applying median filter.


Again applying median filter
Again applying median filter
CONCLUSION:

Image denoising is one of the most fundamental difficulties in image processing and computer
vision, with the goal of estimating the original image by suppressing noise from a noise-
contaminated version of the image. Image noise can be created by a variety of intrinsic (sensor)
and extrinsic (environment) factors that are difficult to prevent in real-world settings. As a result,
picture denoising is useful in a variety of applications, including image restoration, visual
tracking, image registration, image segmentation, and image classification, where recovering the
original image content is critical for good performance. While various methods have been
developed for image denoising, the problem of image noise suppression remains unsolved,
particularly in instances when the photographs were taken in low lighting with a high degree of
noise.

You might also like