SMD-YOLO: An efficient and lightweight detection method for mask wearing status during the COVID-19 pandemic

Abstract

Background and objective: At present, the COVID-19 epidemic is still spreading worldwide and wearing a mask in public areas is an effective way to prevent the spread of the respiratory virus. Although there are many deep learning methods used for detecting the face masks, there are few lightweight detectors having a good effect on small or medium-size face masks detection in the complicated environments.

Methods: In this work we propose an efficient and lightweight detection method based on YOLOv4-tiny, and a face mask detection and monitoring system for mask wearing status. Two feasible improvement strategies, network structure optimization and K-means++ clustering algorithm, are utilized for improving the detection accuracy on the premise of ensuring the real-time face masks recognition. Particularly, the improved residual module and cross fusion module are designed to aim at extracting the features of small or medium-size targets effectively. Moreover, the enhanced dual attention mechanism and the improved spatial pyramid pooling module are employed for merging sufficiently the deep and shallow semantic information and expanding the receptive field. Afterwards, the detection accuracy is compensated through the combination of activation functions. Finally, the depthwise separable convolution module is used to reduce the quantity of parameters and improve the detection efficiency. Our proposed detector is evaluated on a public face mask dataset, and an ablation experiment is also provided to verify the effectiveness of our proposed model, which is compared with the state-of-the-art (SOTA) models as well.

Results: Our proposed detector increases the AP (average precision) values in each category of the public face mask dataset compared with the original YOLOv4-tiny. The mAP (mean average precision) is improved by 4.56% and the speed reaches 92.81 FPS. Meanwhile, the quantity of parameters and the FLOPs (floating-point operations) are reduced by 1/3, 16.48%, respectively.

Conclusions: The proposed detector achieves better overall detection performance compared with other SOTA detectors for real-time mask detection, demonstrated the superiority with both theoretical value and practical significance. The developed system also brings greater flexibility to the application of face mask detection in hospitals, campuses, communities, etc.

Keywords: COVID-19; Computer vision; Face mask detection; Object detection; YOLO.

Fig. 1

Workflow of the face mask detection and monitoring system.

Fig. 2

Structure scheme diagram of SMD-YOLO detector.

Fig. 3

Structure diagrams of partial residual module. (a) The first 3 × 3 convolution layer of original CSPBlock. (b) Design of an enhanced module based on (a) in SMD-YOLO. (c) Design of a lightweight module based on (a) in SMD-YOLO.

Fig. 4

Design scheme of a cross fusion module.

Fig. 5

Structure diagram of the improved prediction network.

Fig. 6

Schematic diagram of an EDAM module.

Fig. 7

Structure diagram of a SPPDAM module.

Fig. 8

Relationship of AvgIOU and the quantity of anchor boxes.

Fig. 9

Distribution of bounding boxes and anchor boxes.

Fig. 10

Image samples of data set for different class labels. (a) With mask (WM). (b) Without mask (WOM). (c) Mask worn incorrectly (WMI) and Mask area (MA).

Fig. 11

Visualization analysis diagrams of the face mask detection dataset. (a) Categorical distribution. (b) Normalized distribution of center points.

Fig. 12

Comparison of P-R curves of different categories.

Fig. 13

Comparison of two detectors for visualization of the heat maps. (a)Original images of test set. (b)Baseline. (c)Our proposed SMD-YOLO.

Fig. 14

Normalization analysis of multiple indicators.

Fig. 15

The detection results of SMD-YOLO detector with real-world face masks wearing status under the various environments. (a)(b)(c) On medium individual faces with a distinctive mask. (d)(e) At the entrance of the hospitals. (f)(g) When small queuing or registering in a small scale at the hospitals. (h)(i) In the densely crowded public areas indoors and outdoors. (j)(k) In the nighttime environment. (l)(m) Inside the hospital corridor from the camera of monitor.