darglein/ADOP

Health v2 · maintenance only

Adoption not part of the score

2026 stars · 192 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

ADOP is a point-based differentiable neural rendering pipeline for scene refinement and novel view synthesis, implemented in C++/CUDA with PyTorch. It renders large point clouds in real time and jointly optimizes camera parameters, point positions/colors, and neural shading networks end-to-end.

Use cases

• novel view synthesis from photos
• refine a point cloud reconstruction with a differentiable renderer
• render 100M+ point scenes in real time
• optimize camera poses and intrinsics jointly with rendering
• reproduce the ADOP paper from ACM TOG 2022

When to choose

• you have an initial point cloud plus camera estimates and want photorealistic novel views
• you need a differentiable point-based renderer for research
• you want real-time rendering of very large point clouds

When to avoid

• you want a maintained production library rather than research code
• you need a plug-and-play NeRF/Gaussian splatting tool with broad community support
• you lack a CUDA-capable GPU

Facets

library · maturity maintenance

machine-learning deep-learning graphics image-processing gpu-computing computer-vision graphics deep-learning machine-learning cpp python neural-rendering novel-view-synthesis differentiable-rendering point-cloud radiance-fields computer-graphics-research linux gpu cuda

Member repositories

For agents

markdown · JSON · MCP: product_card(name="darglein/ADOP")