# IMAGE PROCESSING NUMERICAL QUESTIONS

Click to edit Master subtitle style

5/24/12

Problem 8.4
Consider an 8-pixel line of intensity data, {108,139,135,244,172,173,56,99}. If it is quantized with IGS quantization with 4 bit accuracy, compute the rms error and rms signal to noise ratios for the quantized data.

5/24/12

5/24/12

5/24/12

• What is the maximum compression that can be expected? The maximum compression with Huffman coding is C = 8/5.3 bits/pixel entropy is to be Huffman coded.Problem 8.3 =1. 5/24/12 • .5 A 1024*1024 8 bit image with 5.509.

For instance. 5/24/12 • .• If a greater level of lossless compression is required. compute the differences between adjacent pixels and Huffman code them. what else can be done? One possibility is to eliminate spatial redundancies in the image before Huffman encoding.

Problem 8.9 Consider the simple 4*8. 8-bit image: 21 212195169 243 243 243 21212195169 243 243 243 21212195169 243 243 243 21212195169 243 243 243 5/24/12 .

125 0.375 Probability 5/24/12 .Compute the entropy of the image • Total number of pixel = 32 Grey level R1 R2 R3 R4 21 95 169 243 Value 0.375 0.125 0.

Compression: 5/24/12 .Compress the image using Huffman Coding. and compute the compression achieved.

What is the entropy of the image when looked at as pairs of pixel? 5/24/12 .

811−1.• The difference between this value and the entropy in first part tells us that a mapping can be created to eliminate (1. Difference image: 21 0 0 74 21 0 0 74 21 0 0 74 21 0 0 74 5/24/12 • 74 74 0 0 74 74 0 0 74 74 0 0 74 74 0 0 .56 bits/pixel of spatial redundancy.25) =0.

leaving only (1.25) = 0.16 bits/pixel.The entropy of the difference image: Difference removes most of the spatial redundancy. 5/24/12 .41−1.

If you assume 7-bit ASCII.19 Use the LZW coding algorithm to encode the ASCII string “aaaaaaaaaaa”.Problem 8. the first 128 locations are all that are needed. • Assume that the first 256 codes in the starting dictionary are the ASCII codes. the ASCII ”a” corresponds to location 97. 5/24/12 . In either case.

.5/24/12 The encoded output is : 97 256 257 258 97.

Problem (10.2) •   5/24/12 1515 .

2) • mask 5/24/12 1616 .Solution (10.

Solution (10. Conversely. the response would be a +2 when a mask is centered on a one-pixel gap in a 3pixel segment oriented in the direction favored by that mask.2) Cond… • Each mask would yield a value of 0 when centered on a pixel of an unbroken 3-pixel segment oriented in the direction favored by that mask. 1717 • 5/24/12 .

Problem (10.5)  Suppose that we compute the gradient magnitude of each of the these models Using the Prewitt operator . 5/24/12 1818 . sketch what a horizontal profile through the center of each gradient image would look like.

Solution (10.5) •   -1 -1 -1 0 0 0 1 1 1 5/24/12 1919 .

Cond(10.5)… 5/24/12 2020 .

Problem(10.22) •   5/24/12 2121 .

22) Y=ax+b b   -b/a Normal form xcosθ + ysinθ = ρ 5/24/12 2222 .Solution (10.

22) •   5/24/12 2323 .Solution (10.

5/24/12 2424 .39) • Segment the image shown by using the split and merge procedure .Problem(10. Show the quad tree corresponding to the your segmentation.

39) •   5/24/12 2525 .Solution (10.

Solution (10.39) 5/24/12 2626 .

49) •   5/24/12 2727 .Problem(10.

49)-(a) • Determine the maximum value of K that will guarantee that the blur from motion does not exceed 1 pixel.Problem(10. 5/24/12 2828 .

098 cm/pixel.Solution (10. Thus.5 cm long. the distance between pixels 25/256=0. Because the imaging device is square (256 ×256 elements) the camera looks at an area that is 25cm×25cm. assuming no optical distortions.49)-(a)   It is given that 10% of the image area in the horizontal direction is occupied by a bullet that is 2. 5/24/12 2929 .

the bullet will travel 100. (1.000/0. At this speed.98 = 1. K ≤ 9. 3030 5/24/12 .(a))     The maximum speed of the bullet 1000m/sec = 100. Here it is required that the bullet should not travel more than one pixel during exposure.02×10^6 pixels/sec) ×K sec ≤ 1 pixel.000 cm/sec. So.02×10^6 pixels/sec. That is.Solution Cond…(49.8×10−7 sec.

Problem(10.49)-(b) • Determine the minimum number of frames per second that would have to be acquired in order to calculate the speed of the bullet. 5/24/12 3131 .

5/24/12 3232 .49)-(b) • The frame rate must be fast enough to capture at least two images of the bullet in successive frames so that the speed can be computed.Solution (10.

g. Although these cases could be handled with some processing (e.Solution Cond…(49. leading and trailing edges. as in the case of A and E. and so forth) 3333 • 5/24/12 . In the other cases we get partial bullets..(b)) • If the frame rate is set so that the bullet cannot travel a distance longer (between successive frames) than one half the width of the image. by determining size.

(b))   The length of the bullet in pixels is (2. Because the bullet travels at a maximum speed of 1. One half of the image frame is 128 pixels.Solution Cond…(49. so the maximum travel distance allowed is 102 pixels. The minimum frame rate 1.098cm/pixel) ≈ 26 pixels. 3434 5/24/12 .02×106 pixels/sec.5cm)/ (0.02×106/102 = 106frames/sec.

Starting at different locations in this set does not change the structure of the circular sequence. 5/24/12 . Even if the starting point is not unique.1 Show that redefining the starting point of a chain code so that the resulting sequence of a no. this method would still give a unique sequence. The code for a boundary that is traced in a consistent manner (e.. For example.g.Problem 11. Solution: The key to this problem is to recognize that the value of every element in a chain code is relative to the value of its predecessor. Selecting the smallest integer as the starting point simply identifies the same point in the sequence. but they all yield the same smallest integer 010101. the sequence 101010 has three possible starting points. clockwise) is a unique circular set of numbers. forms an integer of min magnitude makes the code independent of the initial starting point on the boundary .

6 Find an expression for the signature of the boundary of an equilateral triangle and plot the signature.Problem 11. Solution: we see that the distance from the origin to the triangle is given by r(μ) =D0/cos μ 0± · μ < 60± =D0/cos(120± ¡ μ) < 120± μ < 180± μ < 240± μ < 5/24/12 300± =D0/cos(180± ¡ μ) =D0/cos(240± ¡ μ) =D0/cos(300± ¡ μ) 120± · 180± · 240± · 60± · μ .

and D = D0= cos(60±) = 2D0. 5/24/12 .whereD0 is the perpendicular distance from the origin to one of the sides of the triangle. Once the coordinates of the vertices of the triangle are given. determining the equation of each straight line is a simple problem. and D0 (which is the same for the three straight lines) follows from elementary geometry.

) A square Solution: 5/24/12 .) A circle b.7: Draw the medial axis of a.Problem 11.

16 Obtain the gray level co-occurrence matrix of 5*5 image composed of a checkerboard of alternating 0’s and 1’s if The position operator p is defined as “one pixel to the right” Solution: The image is 01010 10101 01010 10101 01010 5/24/12 .Problem 11.

Similarly. and a22 = 0 The total number of pixels satisfying the predicate P is 20. a11 = 0. Element a11 is the number of pixels valued 0 located one pixel to the right of a 0. a21 = 10.Let z1 = 0 and z2 = 1. a12 = 10. By inspection. Since there are only two gray levels the matrix A is of order 2x2 . so C= 0 1/2 1/2 0 5/24/12 .