Deep near-light photometric stereo

This is the project page for our ECCV 2020 paper: "Deep near-light photometric stereo for spatially varying reflectances" by Hiroaki Santo, Michael Waechter, and Yasuyuki Matsushita. [paper]

If you use our paper or code for research purposes, please cite our paper:

@inproceedings{santo2020deep,
title = {Deep near-light photometric stereo for spatially varying reflectances},
booktitle = {European Conference on Computer Vision (ECCV)},
author = {Hiroaki Santo, Michael Waechter, Yasuyuki Matsushita},
year = {2020},
}

Environment

• CUDA 9.0 (CuDNN 7)
• PyTorch (1.1.0)

Dataset

• CAN: https://drive.google.com/file/d/1-1s1njP7M0arN89j45b1Q0SnIZCKwMcB/
• CUP: https://drive.google.com/file/d/1-3MN9UhYUsw4lWyPObBcD6UjpniQVfFd/
• TURTLE: https://drive.google.com/file/d/1-3sBVBfP1k4BL8iMyMVXld3qzJ75Nhde/

Please download and extract them to PATH/TO/DATASET.

How to run

python solve.py --dataset_path PATH/TO/DATASET --obj_name {CAN/CUP/TURTLE} --output_path PATH/TO/OUTPUT

In PATH/TO/OUTPUT, the estimated normal and depth maps are stored as a npz file.