A comparative study of advanced technologies and methods in hand gesture analysis and recognition systems

Hand gesture recognition is progressively becoming a mature technology as a result of significant investment in human–machine interaction. The requirements of human–machine interaction are becoming increasingly genuine for non-verbal communication. In this study, we have analyzed and summarized past studies on non-vision (e.g., data glove-based sensor technology) and vision-based gesture recognition. Several tools and technologies available to date for gesture recognition, including hidden Markov model, finite state machine, color modeling, naive Bayes classifier, deep neural networks, histogram-based features, and fuzzy clustering, have been examined in this study. We have reviewed studies on visual gesture identification based on static and dynamic motions. In the literature, the methodologies presented in gesture recognition have been appropriately divided into the phases of detection, tracking, and recognition, with the various algorithms at each stage developed and contrasted. The purpose of this study is to review prospective technologies, methods, and research outcomes, as well as to analyze the benefits and challenges of various hand gesture detection algorithms, in order to contribute to future research. • Focusing on hand gesture recognition (HGR) for human–machine interaction systems. • Presenting a comprehensive review of emerging technologies for HGR. • Analyzing and summarizing prior studies on vision-based and sensor-based approaches. • Introducing prospective methodologies for developing real-time HGR systems. • Providing a comparison of existing methods for detection, tracking, and recognition.