Submitted by Nicolas Jourdan.

Submission data

Full name	GANet reference submission for robust vision challenge
Description	GANet reference submission as baseline for the stereo benchmark of the robust vision challenge 2020. All method credits go to the original author (Zhang et al.) We are formulating traditional geometric and optimization of stereo into deep neural networks.
Parameters	Trained on Middleburry, KITTI, ETH3D from the KITTI checkpoint made available in the GANet repository on Github by the original authors. Frequency of sampling was adapted to the dataset size. Test images scaled to next multiple of 48.
Publication title	GA-Net: Guided Aggregation Net for End- to-end Stereo Matching
Publication authors	Zhang, Feihu and Prisacariu, Victor and Yang, Ruigang and Torr, Philip HS
Publication venue	CVPR 2019
Publication URL	https://github.com/feihuzhang/GANet
Programming language(s)	Python (PyTorch), CUDA
Hardware	Intel Core I9-9920X, Nvidia Quadro RTX 8000
Website	https://github.com/feihuzhang/GANet
Source code or download URL	https://github.com/feihuzhang/GANet
Submission creation date	17 Jul, 2020
Last edited	17 Jul, 2020

High-res multi-view results

Low-res many-view results

Low-res two-view results

SLAM results