Skip to content
GeoNet: Geometric Neural Network for Joint Depth and Surface Normal Estimation
Python MATLAB Limbo
Branch: master
Clone or download
Latest commit 68d2ba9 Jul 12, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
eval file added Jul 12, 2019
README.md readme Jul 12, 2019
code.py
readme,txt

README.md

GeoNet

GeoNet: Geometric Neural Network for Joint Depth and Surface Normal Estimation

GeoNet++: Interative Geometric Neural network with Edge-aware Refinement Joint Depth and Surface Normal Estimation. (In preparation)

Setup

Requirement

Required python libraries: Tensorflow (>=1.2) + Scipy + Numpy + Scipy + OpenCV.

Tested in Ubuntu + Intel i7 CPU + Nvidia Titan X (Pascal) with Cuda (>=8.0) and CuDNN (>=5.0). CPU mode should also work with minor changes.

Inference

  1. Download pretrained model data, initialization model, and trained model from "https://drive.google.com/open?id=1o2t8735acVf2cLSCS6URkNViOB7mdb-Q"

  2. tar xvzf GeoNet.tar.gz

  3. Merge files into the Repo according to the file name.

  4. Run 'code.py'

  5. Evaluation: cd eval & run 'test_depth.m' for depth evaluation and run 'test_norm.py' for normal evaluation.

Training

Training code already included. Detailed data preparation & running procedure will be updated soon.

You can’t perform that action at this time.