A Fine-grained Detector of Face Mask Wearing Status Based on Improved YOLOX

FineFM Dataset

1. Overview

The Fine-grained Face Mask (FineFM) dataset contains a total of 16,955 images with four classes of face mask wearing conditions.

2. Statistics

1) Train set and validation set

	Train	Val	Sum
Image Number	12,131	1,733	13,864
mask	4,239	615	4,854
face	6,082	793	6,875
mask_chin	5,638	795	6,433
mask_mouth_chin	3,211	447	3,658

2) Test set

	T_easy	T_mid	T_hard	Sum
Image Number	1,600	1,135	356	3,091
mask	400	636	1,814	2,850
face	400	628	337	1,365
mask_chin	400	675	223	1,298
mask_mouth_chin	400	667	326	1,393

3) Comparison of class balance

3. Download

• Google Drive

• OneDrive

• BaiduDisk with extraction code: g6kz

The proposed ECA_YOLOX-S model

1. Network structure

2. Visual results obtained by ECA_YOLOX-S

3. Download

• Google Drive

• BaiduDisk with extraction code: vtxt

Citation

If you think this dataset is helpful, please feel free to leave a star⭐️ and cite our paper:

@article{xiao2023fine,
  title={A Fine-grained Detector of Face Mask Wearing Status Based on Improved YOLOX},
  author={Xiao, Hongli and Wang, Bingshu and Zheng, Jiangbin and Liu, Licheng and Chen, CL Philip},
  journal={IEEE Transactions on Artificial Intelligence},
  year={2023},
  publisher={IEEE}
}