Usage of KINECT to detect walking problems of elder people

• Autor(a) principal: Jesus, Pedro Alexandre Lopes de
• Data de Publicação: 2017
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.6/7838
• topic: Data Normalization; J2k; Kinect; Movement Analysis; Movement Feature Selection; Neural Network; Skeleton Feature Extraction; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática

The dissertation addresses the problem in analyzing and detecting the lack of mobility in the elder population, while enabling a rapid intervention by specialized and able people that can help improve their movement quality. In order to achieve a solution to the problem above, we used the motion detection and gestures device named Kinect, developed by Microsoft, which allowed the recording of all the movements performed by a set of randomly selected people, for the purpose of obtaining data that will allow us to further analysis and classification of motor skills on every person. Additionally, it was necessary to create an application that can extract relevant information generated by the video Kinect and treat this information in order to carry out the person movements classification. Thus, the application is divided into three main steps: -Obtaining the XYZ coordinates, for a relevant set of bones of the person skeleton, in all the recorded frames as well as the total duration of the movement itself; -Extracted data treatment and standardization, so it can later be used in the classifier; -Creation of a classifier using neural networks methodology, which uses the standardized data in order to classify the person movement, according to its quality (existence of a mobility deficit or not). Throughout the dissertation will be described every step of the development process, since the proposed solution design until the code of the developed classes.