Abstract:The detection and tracking of gesture targets is an important aspect in the dynamic gesture recognition. To meet the accuracy and speed requirements of human-computer interaction for dynamic gesture recognition, this study explores long-term gesture recognition under monocular RGB cameras. This study uses an improved particle swarm optimization algorithm as the feature extraction method, and introduces a mixed Gaussian model and kernel correlation filtering to complete gesture detection and tracking. And it has constructed a dynamic gesture tracking model on the ground of kernel correlation filtering. The experimental results show that the skin color based gesture detection algorithm has the minimum average relative error value of 0.321 on different datasets, with accuracy and recall rates higher than 0.8. The maximum correlation coefficient R-squared value is 0.823, and the detection speed reaches 36.32 frames per second. And this detection method has high repeatability on different datasets, with better detection accuracy for different gesture targets. The F1 value of the gesture tracking model has the largest area of the receiver operation characteristic curve, and the two error values of the model are small, resulting in better gesture tracking performance. In human-computer interaction systems, the detection accuracy and target rejection rate of this method have been significantly improved, and the subjective evaluation of the interaction system by the subjects is relatively high, resulting in good application effects. This study enriches the theoretical foundation of dynamic gesture detection and tracking technology, and improves the quality level of gesture tracking in the field of human-computer interaction. This helps to expand the application scope of human-computer interaction.

A Method of Hand Gesture Segmentation and Tracking with Appearance Based on Probability Model

Dynamic hand gesture recognition using hidden Markov models

Hand Gesture Contour Tracking Based on Skin Color Probability and State Estimation Model.

A real-time hand gesture recognition method

2D Motion Detection Bounded Hand 3D Trajectory Tracking and Gesture Recognition under Complex Background

Motionobject segmentation algorithm based on spatial-temporal information

Tracking Multiple People under Occlusion and Across Cameras Using Probabilistic Models

Real-time Dynamic Gesture Recognition and Hand Servo Tracking Using PTZ Camera

Tracking human under occlusion based on kernel density estimation

Based on network probabilistic graph human pose segmentation algorithm

Real-time Detection and Tracking Algorithm Based on the Color and Character of Human Face

Random Walks Colour Histogram Modification for Human Tracking.

Research and Realization on Moving Object Detection and Tracking Methods of Human

Motion-Based Segmentation Scheme to Feature Extraction of Hand Gestures

Gesture Segmentation Based on a Two-Phase Estimation of Distribution Algorithm.

Vision-based Interpretation of Hand Gestures by Modeling Appearance Changes in Image Sequences.

Design of Human–Computer Interaction Gesture Tracking Model Based on Improved PSO and KCF Algorithms

Human body motion tracking based on Bayesian segmentation and grey prediction

An Interactive Image Segmentation Method in Hand Gesture Recognition

Real-Time Hand Locating By Monocular Vision

Motion Objects Segmentation Using a New Level Set Based Method