Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/naoto0804/synshadow
Learning from Synthetic Shadows for Shadow Detection and Removal [Inoue and Yamasaki, IEEE TCSVT 2021].
https://github.com/naoto0804/synshadow
pytorch shadow-detection shadow-removal
Last synced: 9 days ago
JSON representation
Learning from Synthetic Shadows for Shadow Detection and Removal [Inoue and Yamasaki, IEEE TCSVT 2021].
- Host: GitHub
- URL: https://github.com/naoto0804/synshadow
- Owner: naoto0804
- License: mit
- Created: 2020-12-18T05:56:44.000Z (about 4 years ago)
- Default Branch: main
- Last Pushed: 2024-03-16T04:51:06.000Z (9 months ago)
- Last Synced: 2024-03-16T05:29:39.106Z (9 months ago)
- Topics: pytorch, shadow-detection, shadow-removal
- Language: Python
- Homepage:
- Size: 1.01 MB
- Stars: 78
- Watchers: 5
- Forks: 13
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Learning from Synthetic Shadows for Shadow Detection and Removal (IEEE TCSVT 2021)
## Overview
This repo is for the paper "[Learning from Synthetic Shadows for Shadow Detection and Removal](https://arxiv.org/abs/2101.01713)". We present SynShadow, a novel large-scale synthetic shadow/shadow-free/matte image triplets dataset and pipeline to synthesize it. We further show how to use SynShadow for robust and efficient shadow detection and removal.
In this repo, we provide
- SynShadow dataset: `./datasets`
- [SP+M](https://arxiv.org/abs/1908.08628) implementation: `./src`
- Trained models and results: below
If you find this code or dataset useful for your research, please cite our paper:
```
@article{inoue2021learning,
title={{Learning from Synthetic Shadows for Shadow Detection and Removal}},
author={Inoue, Naoto and Yamasaki, Toshihiko},
journal={IEEE Transactions on Circuits and Systems for Video Technology},
year={2021},
volume={31},
number={11},
pages={4187-4197},
doi={10.1109/TCSVT.2020.3047977}
}
```
## Trained Models and Results
We provide the models for shadow detection and removal for convenience. Downloaded models should be placed under `./checkpoints`.
### Shadow Detection
ALl the results are in 480x640. BER is reported for 480x640 images. Below are results evaluated on ISTD test set. DSDNet++ is a modified variant of [DSDNet](https://openaccess.thecvf.com/content_CVPR_2019/html/Zheng_Distraction-Aware_Shadow_Detection_CVPR_2019_paper.html).
| Model | Train | BER | |
| :-: | :-: | :-: | :-: |
|DSDNet++|SynShadow|2.74|[results](https://drive.google.com/file/d/1jhVXHSHj4teacLTK_I-Ep4WNDxId-_tp/view?usp=drive_link) / [weights](https://drive.google.com/file/d/13IhP8d5Tb7ZM3kgu0HmgAvSxAS0KMtXZ/view?usp=drive_link)|
|DSDNet++|SynShadow->ISTD|1.09|[results](https://drive.google.com/file/d/14Ni8fEoQJNj1xz2e_JD1zUKSWwuuwcxc/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1CQ6KjDlKesdb86tJnVcX79B3br_w5jHt/view?usp=drive_link)|
|BDRAR|SynShadow|2.74|[results](https://drive.google.com/file/d/1IHTf0901Mzm5a2G1sROfAyJYoddFNVbF/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1y-VQBoclNzPsfBZovMwpymlVb4I3ljVy/view?usp=drive_link)|
|BDRAR|SynShadow->ISTD|1.10|[results](https://drive.google.com/file/d/1Pr7iAxPsjdZMwwEv7aZ54eiw7GMVoEsS/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1pLZr8uRfnPVkUzI3f7Yx0smcNN2W0l0N/view?usp=drive_link)|
### Shadow Removal
ALl the results are in 480x640. For the pre-trained weights, we only provide SP+M weights, since this repository has full implementation of it. RMSE is reported for 480x640 images.
Model: [SP+M](https://arxiv.org/abs/1908.08628)
| Train | Test | RMSE | |
| :-: | :-: | :-: | :-: |
|SynShadow|ISTD+|4.9|[results](https://drive.google.com/file/d/1r4uM280UIh1XLilHJoPWWiRQ6vBy0fmZ/view?usp=drive_link) / [weights](https://drive.google.com/file/d/10wLm8nVXdmXeBiSO5HfnGT3WNY6AvO8m/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/173WYHwKCV-Rs-sECEQ2Ia_xffY8d_8lS/view?usp=drive_link)|
|SynShadow->ISTD+|ISTD+|4.0|[results](https://drive.google.com/file/d/17R1byVIJ10dRduy_ZforioH2GQTQCenZ/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1dp28xws9TJ7gEyBeiSc0zLCQ09ALptTz/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/1qPi4dA-0TJvYReE2oCoC9zDyGMyxz8kN/view?usp=drive_link)|
|SynShadow|SRD+|5.7|[results](https://drive.google.com/file/d/1-hpQzJ7gGiwNcllVgFqBs9O-yFR1S1YN/view?usp=drive_link) / [weights](https://drive.google.com/file/d/10wLm8nVXdmXeBiSO5HfnGT3WNY6AvO8m/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/1zWWIqZV6P09TTuTIfdkz2NlJZ-5z9Bc-/view?usp=drive_link)|
|SynShadow->SRD+|SRD+|5.2|[results](https://drive.google.com/file/d/13e0AESmIbQxjjvZP9mXjdRYZz6ZQO0s3/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1h_f7O8D8eqNClQ-uKD0MxyGiwhOhIJSi/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/1MoQPbudiFmHJ1hdcwYOjkdvH_ngRceGb/view?usp=drive_link)|
|SynShadow|USR|-|[results](https://drive.google.com/file/d/1BM2m9HHXnFN6JLmcD85LLySe3j6E2cON/view?usp=drive_link) / [weights](https://drive.google.com/file/d/10wLm8nVXdmXeBiSO5HfnGT3WNY6AvO8m/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/1MgH0WWHdzKCWnrMNAMcPNtI6tmBbYhMr/view?usp=drive_link)|
Model: [DHAN](https://arxiv.org/abs/1911.08718)
| Train | Test | RMSE | |
| :-: | :-: | :-: | :-: |
|SynShadow->ISTD+|ISTD+|4.6|[results](https://drive.google.com/file/d/1ci50ArxqLq5VygyVhvTJH1uUmhuhmYKp/view?usp=drive_link)|
|SynShadow->SRD+|SRD+|6.6|[results](https://drive.google.com/file/d/1oiI_Vj5IFMrlT67XMsAF8cHmrxjpVtUY/view?usp=drive_link)|
|SynShadow|USR|-|[results](https://drive.google.com/file/d/1M4CAUvfCdvFnQ_sXuH8ymjhRtaIOGq8e/view?usp=drive_link)|
Note: we have accidentally removed some files and cannot provide some results.