Author(s)

GuoLin Tan^1,2*, HengHui Xiao¹, FangJiong Chen², YanHua Li¹, Ning Lin¹, QunLi Xiao¹

Affiliation(s)

¹The Central Research Institute of Guangdong Communications Services Company Limited, Guangzhou 510630, Guangdong, China.

²School of Electronic and Information Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China.

Corresponding Author

GuoLin Tan

ABSTRACT

In the field of computer vision, the most of applications rely on large training datasets, especially tasks such as object detection, which require a large amount of high-quality labeled datasets. However, in real-world tasks, many object detection tasks are small probability events, such as fighting, climbing over railings, wading, etc. These small probability events face difficulties in obtaining labeled datasets with high cost. In order to solve the problem of detecting people wading, this paper proposes a new low-cost and efficient method for detecting personnel wading in small probability event scenarios. This method uses water semantic segmentation and personnel object detection algorithms to segment the water area and detect people positions respectively, and then combines the water area and the target position information to accurately determine whether the person is wading. In addition to personnel wading detection, this method can also be applied to other similar scenarios, such as parks, scenic spots, factories, and other places where people's activities need to be monitored. By adjusting the algorithm and rules, this method can achieve fast and efficient recognition of human behavior in different scenarios.

KEYWORDS

Computer vision; Small probability events; Image; Object detection; Semantic segmentation

CITE THIS PAPER

GuoLin Tan, HengHui Xiao, FangJiong Chen, YanHua Li, Ning Lin, QunLi Xiao. Method for detecting wading in small probability event scenarios. Journal of Computer Science and Electrical Engineering. 2024, 6(4): 20-24. DOI: 10.61784/jcsee3022.

REFERENCES

[1] Liu H, Frej M B H, Wen B. A novel method for recognition, localization, and alarming to prevent swimmers from drowning//2019 IEEE Cloud Summit. IEEE, 2019: 65-71.

[2] Liu T, He X, He L, et al. A video drowning detection device based on underwater computer vision. IET image processing, 2023, 17(6): 1905-1918.

[3] Yang R, Wang K, Yang L. An Improved YOLOv5 Algorithm for Drowning Detection in the Indoor Swimming Pool. Applied Sciences, 2023, 14(1): 200.

[4] Li D, Yu L, Jin W, et al. An improved detection method of human target at sea based on Yolov3//2021 IEEE International Conference on Consumer Electronics and Computer Engineering (ICCECE). IEEE, 2021: 100-103.

[5] Handalage U, Nikapotha N, Subasinghe C, et al. Computer vision enabled drowning detection system//2021 3rd International Conference on Advancements in Computing (ICAC). IEEE, 2021: 240-245.

[6] He Q, Mei Z, Zhang H, et al. Automatic real-time detection of infant drowning using YOLOv5 and faster R-CNN models based on video surveillance. Journal of Social Computing, 2023, 4(1): 62-73.

[7] Shatnawi M, Albreiki F, Alkhoori A, et al. Deep learning and vision-based early drowning detection. Information, 2023, 14(1): 52.

[8] Alqahtani A, Alsubai S, Sha M, et al. Falling and drowning detection framework using smartphone sensors. Computational intelligence and neuroscience, 2022, 2022(1): 6468870.

[9] Kozlov E, Gibadullin R. Prerequisites for developing the computer vision system for drowning detection//E3S Web of Conferences. EDP Sciences, 2024, 474: 02031.

[10] Jalalifar S, Belford A, Erfani E, et al. Enhancing water safety: exploring recent technological approaches for drowning detection. Sensors, 2024, 24(2): 331.

[11] Zhou B, Zhao H, Puig X, et al. Scene parsing through ade20k dataset//Proceedings of the IEEE conference on computer vision and pattern recognition. 2017: 633-641.

[12] Zhou B, Zhao H, Puig X, et al. Semantic understanding of scenes through the ade20k dataset. International Journal of Computer Vision, 2019, 127: 302-321.

[13] Lin T Y, Maire M, Belongie S, et al. Microsoft coco: Common objects in context//Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V 13. Springer International Publishing, 2014: 740-755.