16-889 HW3 by Sheng-Yu Wang (andrew ID: shengyu2)

Q1.3

grid/ray visualization

grid ray

Q1.4

point sampling

Q1.5

volume rendering

gif depth

Q2.1

See code.

Q2.2

See code.

Q2.3

visualization

mine instructor's

Q3

visualization

mine instructor's

Q4

View dependence implemented. Adding view dependence adds the DOF to the model, and thus improves the expressiveness of the representation (e.g. modeling specular objects). However, a more complex model will have larger sample complexity, so more views will be needed to generalize to novel views. In the provided scene (lego), adding viewing dependence does not visually improve the results drastically.

w/o view dependence w/ view dependence