Graphics & Games · 2018

Image and Mesh Processing

Image filters, dithering, resampling and quadric error mesh simplification from scratch.

The Stanford bunny simplified to progressively fewer triangles

A flower photo quantized to 1, 2, 3, 4 and 8 bits per channel

A flower photo with Floyd-Steinberg dithering at several bit depths

The mandrill test image blurred with increasing kernel sizes

Edge detection applied to a flower photo

The Great Wave downscaled with nearest neighbor, hat and Mitchell filters

A photo of a shed warped with a fish-eye filter

The mandrill test image warped with a fish-eye filter

1 / 9

As part of CSE 163: Advanced Computer Graphics at UCSD in spring 2018, we implemented a range of image and mesh processing techniques.

Image processing

Brightness, contrast, saturation and gamma adjustment.
Quantization and dithering.
Integer convolution for blurring, sharpening and edge detection.
Anti-aliased scaling and shifting.
A non-linear fish-eye filter.

Mesh processing

Phong-based OpenGL rendering.
Edge collapsing.
Fast LOD generation by selecting the edges that minimize quadric error, as described in Garland’s Surface Simplification Using Quadric Error Metrics.