8000 GitHub - ML3DSLL/occupancy_networks: This repository is the project for course Machine Learning for 3D Geometry
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

ML3DSLL/occupancy_networks

 
 

Repository files navigation

Occupancy Networks for Single Image Reconstruction

Example 1 Example 2 Example 3

This repository contains the code for our project for the course 'Machine Learning for 3D Geometry' at TUM [IN2392]. In this work, we had improved the performance of the model in terms of Intersection over Union (IoU) compared to the baseline Occupancy Networks - Learning 3D Reconstruction in Function Space. Additionally, we had reduced the number of parameters to make the training of the model more efficient.

Authors

  1. Yujun Lin
  2. Joong-Won Seo
  3. Yunan Li

Dataset

For our experiments, we use Pix3D as our new dataset which contains images, masks, meshes and camera positions.

Modifications

  1. Replace the backbone ResNet with ConvNeXt
  2. Integrate feature pyramid to enable multi-scale inputs
  3. Incorporate camera pose information

Results

To see the experiment results, please check the final paper [ML3D_Report.pdf] in the repository:

About

This repository is the project for course Machine Learning for 3D Geometry

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 85.9%
  • Python 7.8%
  • C++ 2.1%
  • Cuda 1.8%
  • C 1.2%
  • Cython 0.6%
  • Other 0.6%
0