Posture detection, classifier and correction system for cyclists applying KINECT V2 with neural nets and fuzzy logic

摘要

The development of computer systems that analyze the execution of top athletes has been improving and becoming more meticulous and precise in an attempt to measure each variation that affects the performance of athletes. These complex systems are reserved, due in most part to their high costs and lengthy implementation process, to a few selected groups of athletes, excluding the amateur group, which represents a considerable population. The same happens with cycling, so the purpose of developing a low-cost, and rapid implementation system, is to conduct a larger group of people towards a comfortable riding posture and, consequently, to improve their performance.