Having a number of Webcam-captured video streams, build a Web application that detect, classify, compare, and synchronize hand and arm gestures performed by (human) users. A conceptual model will be created/(re)used in order to express (classes of) gestures, anatomic features, associated actions, etc. A rule-based approach could be adopted – for example, if the "wave" gesture is detected in at least 74% of video feeds exposed by a video-conferencing system, then the conference session will be ended. Also, different statistics of interest will be offered in graphical form and as JSON-LD data. Study An Ontology for Reasoning on Body-based Gestures. Consult also Awesome Streaming and Recognition APIs. Bonus: capturing and exposing useful provenance.
