| dc.description.abstract | Currently, with the increasing evolution of technology, there are countless researches in the area of health aimed at improving the quality of life of patients. Technologies such as Machine Learning, for example, enable automated data analysis extremely fast due to the processing speed of computers, especially with the use of advanced graphics cards. The programs of patients with chronic obstructive pulmonary disease are too dependent on manual analysis, a fact that limits the patients that each professional can attend. Likewise, the data to be analyzed are limited because they are recorded manually. The model described in this work is of an applied nature, since it has the objective of solving a practical problem. The data were analyzed in a qualitative way, through a use analysis, without quantifying results. The objectives of this work fit as an exploratory research, because a new solution will be proposed, which will meet the proposed problem and expand the knowledge inherent to the context that has already been explored. The study has technical bibliographic and experimental procedures. It was implemented a model application that allowed the extraction of movements of keypoints of the body for historical monitoring and evaluation of the patient's progress in a manual or automated way, using the similarity of time series through Euclidean distance and dynamic time warping algorithms. With the model presented, it was possible to determine the percentage of increase of about 75% in the similarity of the movements of exercises during sessions of physiotherapeutic treatment. | en |
