Note: @ Gi) Time: Reg. No. | | | | B.Tech/ M.Tech (Integrated) DEGREE EXAMINATION, MAY 2024 Fourth Semester 21CSE251T — DIGITAL IMAGE PROCESSING (For the candidates admitted from the academic year 2022-2023 onwards) ‘Part - A should be answered in OMR sheet within first 40 minutes and OMR sheet should be handed over to hall invigilator at the end of 40" minute. Part - B and Part - C. should be answered in answer booklet. 3 Hours PART — A (20 x 1 = 20Marks) ‘Answer ALL Questions 1 is additive color models. (@) His (B) RGB (©) CMYK (D) CMY iS . A pixel ‘P” at coordinates (x,y) has neighbors whose coordinates are given (xt, yt), +1, y-1), (xl, y+ 1), (X-1, y-1). This set of pixels is known as (A) Diagonal neighbors (B) 4-neighbors (C)_ 8-neighbors (D) None of the mentioned 3. Which factor is responsible for determining the spatial resolution of an image? (A) Dynamic range (B) Sampling (C) Contrast (D) Quantization 4, Identify the next step in image processing after compression (A) Acquisition (B) Segmentation (C) Enhancement (D) Morphological processing 5. Which of the following transformation is used in Cathode Ray Tube (CRT) devices? (A) Log transformation (B) Power-law transformation (©) Negative transformation (D) Antilog transformation 6. The output of smoothing, linear spatial filtering isa_of the pixels contained in the neighborhood of the filter mask. (A) Sum (B) Product (C) Average (D) Dot product 7. The principle objective of sharpening, to highlight transition is (A) Brightness (B) Composure (C) Pixel density (D) Intensity 8. Histogram equalization is mostly used for (A) Contrast adjustment (B) Blurring (C)_ Image enhancement (D) None of the mentioned Page 1 of4 Max. Marks: 75 Maris BL CO PO por aw 1 ot 2 ea » 1 2 Ie bo 2 ie 2MA4-21CSE251T10. i 12. 13, 14, 15. 16. 17, 18, 19. 20. In Weiner filtering, it is assumed that noise and images are (A) Correlated (B) Different (©) Uncorrelated (D) Homogeneous Major source of noise arise during image : (A). Restoration (B) Enhancement (C) Degradation (D) Acquisition In geometric mean filter when alpha is equal to 0 then it works as (A) Notch filter (B) Parametric filter (C) Bandpass filter (D) Inverse filter The universe of image convolution is (A)_ Image non-convolution (B) Image in convolution (C)_ Image deconvolution (D) Image by convolution ‘Compressed image can be recovered back by (A) Image enhancement (B) Image contrast (C)_ Image decompression (D) Image equalization Digital image with intensity levels in the range [0, L~1] is called (A), K-map (B) Histogram (C) Truth table () Graph Which of the following is the example of lossy compressi (A) JPEG (B) BMP (C) Huffman coding (D) Delta coding Region of interest (Rol) operations is commonly called as (A). Shading correction (B) Masking (C) Dilation (D) None of the mentioned Which of the following is a type of neural network? (A) Decision tree (B) Random forest (C) Convolutional neural network (D) Linear regression Which of the following is an example of a supervised learning problem? (A) Image classification (B) Market segmentation (C) Fraud detection (D) Social network analysis What is the importance of using PCA before the clustering? Choose the most complete answer? (A) Find good features to improve (B) Find the explained variance your clustering score (©) Avoid bad features (D) Find which dimension of data maximize the features variance _____algorithm that propagates errors from nodes of output to input. (A) Backpropagation (B) Front propagation (C) Signal propagation (D) Channel propagation Page? of 4 11 3) 13,0 1 1 3 13380 13 135)0 11 4 16 tot 4 124 Dot 4 aaa 11 «12a 11 5 1243.0 11 5 1345.0 11S 1345.10 11S 13480 ‘22MA4-21CSE2SITPART —B (5 x 8 = 40 Marks) Mats BL CO PO Answer ALL Questions 21, a. Describe the fundamental steps used for processing digital images with § 2 1 12 block diagram. (OR) b. Explain the following relationship between pixels es 41 wp (i) Connectivity and Gi) Distance measures 22. a, Perform histogram equalization on the image below. $6 2 1a 1 3 5 5 [2{1] (OR) b, Consider the segment of the image below V = {0, 1}. Calculate the * 6 2 124 Euclidean distance D4 and D8 between two pixels p and q. Oi [2[3] PRpRy tli fo]3 20 23.a. Explain the concepts of degradation model. #230 13s (OR) b. Derive a wiener filter for image restoration and explain its advantages over § 4 3 135 the inverse filter 24. a, What is the need for image compression? Explain two image compression § 2 4 124 techniques in detail. (OR) b. Explain the following features $2 4 pe (i) Histogram-based features (i) Intensity features 25.a. Describe CNN and demonstrate how it detects objects 82s 1asu0 (OR) b. What is binary classifier and how does SVM work in image classification? @ 4 § Explain it in a neat block diagram. PART — C (1 x 15 = 15 Marks) Marke BL co FO Answer ANY ONE Question 26. Consider the image below and determine the output of pixel (2,2) is. 8 6 2 124 smoothing is performed using the 3x3 neighborhood and all of the filters listed below. Page 3 of 4 22MAS-21CSE2SIT8/8 /)0|7 7/9|5|7 4/6\/8/6 2/0/1/5 0210 Box filter Weighted average filter Median filter Max filter Min filter 27. Design a system for detecting driver drowsiness using image processing techniques, suggest a suitable algorithm for each step. eee Page 4 of 4 22MAG-21CSESIT