Chinese Journal of Liquid Crystals and Displays, Volume. 38, Issue 8, 1139(2023)
Mask wearing detection based on improved YOLOv7
Figures & Tables(10)
Fig. 7. Comparison of Fast-RCNN, YOLOv7 and improved YOLOv7 detection results.
Table 1. Classification and number of targets in the data set
View tableView in ArticleTable 1. Classification and number of targets in the data set
Defect type Training set Validation set Test set Total No-mask 3 112 778 778 3 890 Wearing correctly 2 352 588 588 2 940 Wearing incorrectly 1 920 480 480 2 400
Table 2. Performance index of different improvement methods
View tableView in ArticleTable 2. Performance index of different improvement methods
Model CBAM ConvNeXt SPPCSPCP Precision/% mAP/% YOLOv7 90.2 90.2 YOLOv7-A + 93.4 90.9 YOLOv7-B + + 92.9 91.9 YOLOv7-C + + + 92.4 93.8
Table 3. Comparison of performance indicators of different detection algorithms
View tableView in ArticleTable 3. Comparison of performance indicators of different detection algorithms
Method No-mask/% Wearing
correctly/%
Wearing incorrectly/% mAP/
%
Faster-Rcnn 74.1 70.9 72.2 72.4 YOLOv7 87.4 94.1 89.2 90.2 YOLOv7-A 92.1 95.7 84.8 90.9 YOLOv7-B 94.1 95.6 88.6 91.9 YOLOv7-C 94.2 96.2 90.9 93.8
Tools
Get Citation
Copy Citation Text
Hui-chen FU, Jun-wei GAO, Lu-yang CHE. Mask wearing detection based on improved YOLOv7[J]. Chinese Journal of Liquid Crystals and Displays, 2023, 38(8): 1139
Paper Information
Category: Research Articles
Received: Nov. 8, 2022
Accepted: --
Published Online: Oct. 9, 2023
The Author Email: Jun-wei GAO (qdgao163@163.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20