A PyTorch Library for Accelerating 3D Deep Learning Research

Quickly and accurately render even the largest data.

A geometry-shader-based, global CUDA sorted high-performance 3D Gaussian Splatting rasterizer. Can achieve a 5-10x speedup in rendering compared to the vanialla diff-gaussian-rasterization.

Original reference implementation of "StopThePop: Sorted Gaussian Splatting for View-Consistent Real-time Rendering"

Analysis of georeferenced rasters, vectors and point clouds

Differentiable Point Radiance Fields Rasteriser for Novel View Synthesis

Enterprise-ready Document Analysis with Large Language Models

A software 3d rasterizer for pygame (no other dependencies, e.g. numpy)

HISDAC-ES: Creating historical settlement data for Spain (1900-2020) based on cadastral building footprint data

An improved AHN3 gridded DTM/DSM done as university project for the MSc Geomatics @ TU Delft

Some projects are modified from Chu-Song Chen's class of 3D Computer Vision with Deep Learning Applications at National Taiwan University.

Selected python scripts for geoprocessing using open source geospatial resources

Learning the basics of rendering with PyTorch3D, exploring 3D representations, and practicing constructing simple geometry.

OMTP performance collection

A custom rasterization tool utilizing Pyngine to draw 3D objects.

Implementação de métodos de rasterização - Computação Gráfica

Reconstructed the OpenGL API engine in Python and implemented the pipeline to render 2D and 3D models in the engine.

A 3D python rasterizer that renders full 3D scenes from loaded in .obj files

Reconstructed the OpenGL API engine in Python and implemented the pipeline to render 2D and 3D models in the engine.