It is a pivotal problem for accurate and efficient human body detection in the field of computer vision. However, the complex backgrounds, various body postures, occlusions, shadow and so forth that usually have a negative impact on the performance of human body detection. Besides, the real-time ability of the existing detection algorithms are limited in the practical application. In this paper, with the excellent learning ability, a fast and efficient deep convolution neural network based on the YOLOv2 network is presented for real-time human body detection. It is a 22-layer network that is capable to handle the dataflow in 93.5 fps, fully meets the real-time requirements. In the same time, it achieves 80.27% average precision in the complex natural scene.
CITATION STYLE
Liu, X., Liu, Y., Wang, H., & Li, J. (2020). Real-Time Human Body Detection Based on YOLOv2 Network. In Lecture Notes in Electrical Engineering (Vol. 582, pp. 465–473). Springer. https://doi.org/10.1007/978-981-15-0474-7_44
Mendeley helps you to discover research relevant for your work.