The Detection, Extraction, and Classification of Human Pose in Alzheimer's Patients

Abstract

From a biological perspective, aging is the consequence of the accumulation of extensive molecular and cellular damage over time, which is especially seen in elderly people with neurological diseases such as Alzheimer’s who need to move about or who need to move from one side of a room to another, or simply walk and sit because they may feel either anxious or agitated. In this case, the main objective of this research paper is aimed at identifying the estimation of the “Standing” and/or “Sitting” pose in Alzheimer's patients from images obtained from elderly care centers in the canton of Ambato, Ecuador, which is to be used later in an exploratory analysis related to the categories of wandering, nervous, depressed, disoriented, or bored. We worked with a population of 45 people from both sexes, who were diagnosed with Alzheimer's and whose ages ranged between 75 and 89. The methods used were pose detection, feature extraction, and pose classification. As a result, the physical states of “standing” and “sitting” were identified with the usage of algorithms that facilitated the adequate operation of pose estimation. It is concluded that the procedures and data that were obtained provide key input for future research in the health fields related to the behavioral aspects of patients.