Whether this loss can be used directly to calculate disparity loss

aim-uofa / AdelaiDepth

This repo contains the projects: 'Virtual Normal', 'DiverseDepth', and '3D Scene Shape'. They aim to solve the monocular depth estimation, 3D scene reconstruction from single image problems.

Creative Commons Zero v1.0 Universal

1.06k stars 144 forks source link

Whether this loss can be used directly to calculate disparity loss #38

Closed aiforworlds closed 2 years ago

aiforworlds commented 2 years ago

Hello, thank you for your amazing work!

I want to kown whether (PWNPlanesLoss、EdgeguidedNormalRegressionLoss、MSGIL_NORM_Loss)loss can be used directly to calculate disparity loss, because my dataset is disparity data not depth data.

Thank you for your time Best Regards

guangkaixu commented 2 years ago

Hello @aiforworlds , you can train with gt disparity by providing annotations like: [{"rgb_path": "folder/rgb/0.jpg", "disp_path": "folder/disp/0.png"}, ...] , where disp_path is the path of gt disparity image. The disparity will be converted to depth in multi_dataset.py (https://github.com/aim-uofa/AdelaiDepth/blob/6309e00ae56b8fc74a52431f76d8e18633351c70/LeReS/Train/data/multi_dataset.py#L332).