02098nas a2200229   4500000000100000000000100001008004100002260005400043653001800097653002800115653001700143653002800160653003700188100001900225700001200244700001800256245010900274856015300383300001400536520127600550020004201826       2020                            d  c2020///bSpringer Science and Business Media B.V.10aDeep learning10aArtificial intelligence10aApplications10aFolk dance preservation10aHuman motion analysis from video1 aN. Grammalidis1 aI. Kico1 aF. Liarokapis00aAnalysis of Human Motion Based on AI Technologies: Applications for Safeguarding Folk Dance Performances  uhttps://www.scopus.com/inward/record.uri?eid=2-s2.0-85126113393&doi=10.1007%2f978-3-030-36126-6_35&partnerID=40&md5=8734a7db3ec0534f60f17f09de857137  a321-329, 3 aAnalysis of human motion is an important research area in computer vision with numerous applications. Recent projects, such as EU i-Treasures and TERPSICHORE projects conduct research in this field to improve the capture, analysis and presentation of Intangible Cultural Heritage (ICH) using ICT-based approaches. The final goal is to document these forms of intangible heritage and to capture the associated knowledge in order to safeguard and transmit this information to the next generations. In addition, these approaches can give rise to new services for research, education and cultural tourism. They can also be used by creative industries (e.g. companies performing film, video, TV or VR applications production), as well as by local communities, creating new local development opportunities by promoting local heritage. This paper first reviews some very recent state of the art approaches based on deep learning which can achieve impressive results in recovering human motion (2D or 3D) and structure (skeleton with joints or realistic 3D model of the human body). Based on such approaches, we then propose a dance analysis approach, currently under development in TERPSICHORE project. Preliminary results are presented and, finally, some conclusions are drawn.  a21987246 (ISSN); 9783030361259 (ISBN)