Τεχνικές μηχανικής μάθησης εφαρμοσμένες στην άυλη πολιτιστική κληρονομιά
The aim of the research study and present diploma thesis is the study of various machine learning techniques for the digitization of the intangible cultural heritage. This diploma thesis is structured in two different parts, each of which studies and analyzes a different application and framework on the intangible cultural heritage. In the first section will be presented an alternative approach to choreography summarization. This means that a very small number of key-frames from images are exported in order to represent a whole choreography, thus significantly reducing the processing and storage complexity. In our approach, the problem of the summary of choreography is treated as an approach of unsupervised clustering. In the second section, an educational framework will be presented for analysis and visualization of dance kinesiology based Labanotation and embodied learning concepts. The Kinect sensor is used to extract skeletal data, which are then processed and transformed geometrically. In the sequel they are analyzed based on the Labanotation system to characterize the posture of the human limbs.