16-825 Project 1: Rendering Basics

1. Practicing with Cameras

1.1 360-degree render

360-degree render of the cow mesh

1.2 Dolly zoom

Dolly zoom applied to the cow mesh

2. Practicing with Meshes

2.1 Tetrahedron

The tetrahedron has 4 vertices and 4 triangular faces.

360-degree render of a tetrahedron

2.2 Cube

The cube has 8 vertices and 12 triangular faces (two triangles per side).

360-degree render of a cube

3. Re-texturing a Mesh

I used color1 = [1, 0, 0] (red) at the minimum z-coordinate and color2 = [0, 1, 0] (green) at the maximum z-coordinate, with linear interpolation between them.

Cow mesh with a red-to-green interpolated texture

4. Camera Transformations

Transform 1: Rotate 90° in the xy-plane about z; no relative translation.

Cow after the first relative camera transformation

Transform 2: Keep the identity rotation and translate 7 units along z, moving the camera farther away.

Cow after the second relative camera transformation

Transform 3: Apply a small −0.1 radian rotation in the xz-plane about y; no relative translation.

Cow after the third relative camera transformation

Transform 4: Rotate 90° in the xz-plane about y, then translate by [-3, 0, 3].

Cow after the fourth relative camera transformation

5. Rendering Generic 3D Representations

5.1 RGB-D Point Clouds

The first view, second view, and their union are shown side-by-side (every 10th point sampled).

Point cloud from RGB-D image 1 Point cloud from RGB-D image 2 Combined point cloud from both RGB-D images

5.2 Parametric Functions

A torus and a figure-eight Klein bottle immersion, both sampled as point clouds.

360-degree render of a parametric torus point cloud 360-degree render of a parametric Klein bottle point cloud

5.3 Implicit Surfaces

The torus and ellipsoid below were extracted from voxelized implicit functions with marching cubes. Meshes generally render as continuous, smooth surfaces and can be efficient because triangles fill the space between vertices. Point clouds are simpler to create and combine, but often need many samples and careful point-radius tuning to avoid holes. Mesh extraction adds preprocessing and connectivity storage, while point clouds store only independent positions and features.

360-degree render of an implicit torus mesh 360-degree render of an implicit ellipsoid mesh