Video: body posture estimation system
II Turing Machine Awards for the Best End of Studies Projects (TFE), 2022. Work “Study and development of a body posture estimation system” by Cristina Bolaños Peño, and supervised by María José Santofimia Romero and Félix Jesús Villanueva Molina, from ARCO research group. Award for the best TFM granted by the company Ubótica Technologies.
Summary
In the midst of the digital age, technology must be put at the service of the user, supporting, and even improving, different aspects of their daily lives. In the field of health and care for the elderly, physical rehabilitation is receiving more and more attention due to the impact it has on the quality of life of these people. However, access to professionals is not always possible or, at least, with the desired frequency, so the use of new technologies can be a great enabler for the improvement of physical rehabilitation processes at home. In this sense, the estimation of the user's body posture in real time is key to offering an adequate exhaustive and precise monitoring service of the movements made, being able to verify their correctness. A technological solution of this type will serve not only the end user who participates in the rehabilitation program, but also the health personnel who supervise this rehabilitation process, who will have access to quantitative and precise information on its evolution as well as the possibility to reach more users. In addition, for this solution to be widely accepted, it will be necessary to take into account issues of energy and computational consumption, in such a way that the solution must be able to be deployed in embedded devices, with limited computational resources, thus keeping the energy consumption and price of the final platform. In this document, several methods and embedded devices are analyzed and evaluated to obtain a user body pose estimation system that meets the requirements of low energy demand and computational resources. In addition, the development of a library that allows obtaining said estimation from an image to serve a physical rehabilitation system for the elderly at home will be described.
