Gesture Recognition Analysis Based On Monocular Vision Relying On Human-Computer Interaction Filtering Algorithm

Qin Liu

DOI: https://doi.org/10.1088/1742-6596/2037/1/012097

2021-09-01

Journal of Physics: Conference Series

Abstract:The progress of vision-based gesture recognition technology has introduced a brand new way of human-computer interaction (HCI). However, due to the complexity of feature extraction, gesture recognition based on monocular vision has become a research hotspot and bottleneck. In this paper, static gestures are combined with the CSS feature description to extract the skin-color model based on the HCI filtering algorithm first. Subsequently, the fusion feature is established. Finally, feature recognition is performed based on the HCI filtering algorithm to implement gesture recognition analysis based on monocular vision. The practice has demonstrated that the proposed algorithm has a relatively high recognition rate for complex gestures with the sign language of high local similarity.