Assignment 1 - Chan WangSing

1. Practicing with Cameras

1.1 360-degree Renders

I rendered the cow from multiple viewpoints by keeping the camera distance and elevation fixed while varying the azimuth from 0 to 360 degrees.

1.2 Dolly Zoom

I created the dolly zoom by changing the field of view while adjusting the camera distance so that the cow remains approximately the same size.

2. Practicing with Meshes

2.1 Tetrahedron

I constructed the tetrahedron using 4 vertices and 4 triangular faces.

2.2 Cube

I constructed the cube using 8 vertices and 12 triangular faces. Each of the 6 square sides is represented using two triangles.

3. Re-texturing the Cow

I re-textured the cow by interpolating between red and blue based on the z-coordinate of each vertex. I normalized the z-coordinate using alpha = (z - z_min) / (z_max - z_min).

4. Camera Transformations

Transformation 1

I applied a 90-degree clockwise rotation about the z-axis.

Transformation 2

I applied a 90-degree rotation about the y-axis and compensated for the resulting translation to change the camera viewpoint.

Transformation 3

I moved the camera farther away by using a relative translation of [0, 0, 3].

Transformation 4

I shifted the viewpoint right and slightly upward using a relative translation of [0.5, -0.25, 0].

5. Rendering Generic 3D Representations

5.1 RGB-D Point Clouds

I unprojected the two RGB-D images into 3D point clouds using their depth maps, masks, and cameras. The third point cloud is the union of the points from both views.

From left to right: point cloud from the first image, point cloud from the second image, and the union of both point clouds.

5.2 Parametric Functions

Torus

I generated a torus point cloud by sampling its parametric equations.

Additional Object: Möbius Strip

I generated a Möbius strip by sampling its parametric equations and rendering the resulting points as a point cloud.

5.3 Implicit Surfaces

Torus

I defined the torus using an implicit function and extracted its surface as a triangle mesh using marching cubes.

Mesh vs. Point Cloud

A mesh produces a smoother and more continuous surface because neighboring vertices are connected by faces. A point cloud is simpler and does not require connectivity information, but its visual quality depends on the number and size of the rendered points. Meshes generally provide cleaner surface renderings, while point clouds are easier to construct and manipulate.

Additional Object: Ellipsoid

I defined an ellipsoid using an implicit function and extracted the surface using marching cubes.