No. of Pages: 2

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY
THIRD SEMESTER M.TECH DEGREE EXAMINATION, July 201 8
Branch:

Electronics and Communication Engineering
Stream (5): -
Signal Processing

Course Code & Name:

OIEC 7315: Computer Vision
Answer any twofull questions from each part
Limit answers to the required points.

Max. Marks: 60 Duration: 3 hours

PARTA

1. a. Derive the perspective projection equation. State the assumptions used. 7 b Prove

that parallel lines when imaged, converge at a point. 2

2. Explain in detail the steps for computing scale-invariant feature transform
9
(SIFT).
3. a. Explain the use of Hough transform for line detection. 5
b. Explain Harris corner point detector. 4
PART B

Explain in detail the steps involved in structure from motion (SFM) method for

3D reconstruction. 9

5. a. What is aperture problem in optical flow? 2
b. Derive the Horn-Schunck algorithm for computation of optical flow. 7

6. ഒ. Derive the optical flow constraint equation. 3

b. Prove that optical flow obtained using Lucas-Kanade algorithm is the least

6 squared solution of optical flow constraint equation.

PART C

7. a. Define shape from shading. 3

b. Derive an equation for surface normals posing shape from shading as an
optimization problem. 6

8. ഒ. Explainin detail the steps for object detection using histogram of oriented gradients
(HOG). 8