b. Consider estimating optical flow given two images and derive the brightness constancy
constraint. Given an initial guess for the optical flow vectors, derive a linear system
7 of equations to update the same.

6. a. Define fundamental matrix and essential matrix. 3
b. Explain triangulation and bundle adjustment in reference to structure from motion. 6
PART C

.a. Given reflectance map and a single image, explain a method to obtain surface normals corresponding to
real 3D scene that is imaged. 7

b. Explain the shape from shading problem. Explain any one method to solve this problem.

8. a. Explain how texture and focus play a role in how we perceive shape. Explain how these cues can be
used to reconstruct 3D geometry. 6

b. Mobile cameras use automated face detection. Explain any one face detection method that is currently in use.

6
9, a. Explain in detail the steps for pedestrian detection using histogram of oriented gradients, 6
b. Explain how Eigen faces can be used for face recognition in images. 6