In recent years, we have witnessed the widespread use of so-called “Kinect-like devices,” i.e., devices based on a set of low-cost sensors that allow you to obtain a depth image of the scene being captured.

The high accessibility of these devices, mainly in terms of cost, has facilitated their spread in the field of gesture recognition in numerous applications, both commercial and research-related.

This article will first illustrate the general principles underlying the main techniques used to recognize gestures, exploiting the data obtainable from “Kinect-like” devices. Subsequently, some areas of application will be presented, ranging from the educational-recreational sector to the more scientific one (home automation, robotics, and biomedical engineering).

The appendix lists the main products available on the market and provides a brief comparative analysis. It also describes one of the best-known and most widely used skeletal tracking algorithms, on which most gesture recognition solutions are based.

This is one of the scientific articles published by one or more synbrAIn collaborators and data scientists.