pmh47/dirt
DIRT: a fast differentiable renderer for TensorFlow
This project helps researchers working with computer vision or 3D graphics create and manipulate 3D mesh models directly within the TensorFlow framework. It takes descriptions of 3D objects (like vertex positions and textures) and generates 2D images, while also allowing you to calculate how changes to the 3D model would affect the final image. Researchers can use this to train machine learning models that understand or generate 3D shapes from images.
314 stars. No commits in the last 6 months.
Use this if you need to render 3D mesh geometry into 2D images within TensorFlow and require the ability to compute gradients through all aspects of the 3D scene, including geometry, lighting, and textures.
Not ideal if you don't require differentiable rendering capabilities or are not working within the TensorFlow ecosystem, as it has specific GPU and TensorFlow requirements.
Stars
314
Forks
63
Language
C++
License
MIT
Category
Last pushed
Mar 16, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/pmh47/dirt"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
openspyrit/spyrit
A Python toolbox for deep image reconstruction, with emphasis on single-pixel imaging.
RobotLocomotion/pytorch-dense-correspondence
Code for "Dense Object Nets: Learning Dense Visual Object Descriptors By and For Robotic Manipulation"
Fyusion/LLFF
Code release for Local Light Field Fusion at SIGGRAPH 2019
marrlab/SHAPR_torch
SHAPR: Code for "Capturing Shape Information with Multi-Scale Topological Loss Terms for 3D...
natowi/3D-Reconstruction-with-Deep-Learning-Methods
List of projects for 3d reconstruction