We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
轻量化二维人体骨骼关键点检测算法综述.
- Authors
曾文献; 马月; 李伟光
- Abstract
With the development of mobile devices and embedded devices, higher requirements are put forward for the two-dimen-sional human skeleton key point detection network. Designing a lightweight neural network is an important method to solve the problem of large network parameters and large amount of computation. Firstly, the mainstream methods and lightweight neural networks for two-dimensional human skeletal key point detection based on neural networks were introduced. Then, the lightweight human pose estimation methods based on neural networks in recent years were classified and summarized, and the two-dimensional skeletal key point detection methods were grouped into four categories according to the lightweight way of neural networks: lightweight feature extraction network, deep separable convolution, dense connection mechanism and lightweight bottleneck, and analyzed their advantages and disadvantages and lightweight means. Finally, the common data sets and corresponding evaluation metrics were introduced, and the improved light-weight methods were compared with experimental data. A summary and outlook were given in relation to the current challenges and fu-ture development trends of the research.
- Publication
Science Technology & Engineering, 2022, Vol 22, Issue 16, p6377
- ISSN
1671-1815
- Publication type
Article