Language:
🇺🇸
🇨🇳
«DCL» re-implements the paper Destruction and Construction Learning for Fine-Grained Image Recognition
More training statistics can see:
Differ with other attention-based or part-based fine-classification methods, DCL adds an Destruction Module (Region Confusion Mechanism and Adversarial Learning Network) and Construction Module (Region Align Network) in training, and only use backbone network in infer. Improve the accuracy of the model without affecting the reasoning speed.
Current project implementation is based on JDAI-CV/DCL.
$ pip install -r requirements.txt
- Train
$ CUDA_VISIBLE_DEVICES=0,1,2,3 python tools/train.py -cfg=configs/cub/r50_cub_448_e100_sgd_dcl_5x5_g4.yaml
- Test
$ CUDA_VISIBLE_DEVICES=0,1,2,3 python tools/test.py -cfg=configs/cub/r50_cub_448_e100_sgd_dcl_5x5_g4.yaml
- zhujian - Initial work - zjykzj
@InProceedings{Chen_2019_CVPR,
author = {Chen, Yue and Bai, Yalong and Zhang, Wei and Mei, Tao},
title = {Destruction and Construction Learning for Fine-Grained Image Recognition},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2019}
}
Anyone's participation is welcome! Open an issue or submit PRs.
Small note:
- Git submission specifications should be complied with Conventional Commits
- If versioned, please conform to the Semantic Versioning 2.0.0 specification
- If editing the README, please conform to the standard-readme specification.
Apache License 2.0 © 2021 zjykzj