Abstract:

Obesity is a disease that causes and can be treated with the increase in the amount of body fat. Increased body weight can lead to disorders on the muscle skeletal system. The foot is a structure responsible for the transfer of body weight to the ground. Obesity has been shown to be affected by the anatomical structure of the foot, the bottom contact surface area and pressure changes. This study developed a model that classifies the long short-term memory (LSTM)-based deep nerve network (DSA) and the body mass index (VKI) conditions with the measurement of the foot base pressure values of individuals. First, the pressure values of the 12 different areas of the individual's foot base are measured by a baropodometer. The measurement resulted in 13 different input parameters to be used in the data set of the nervous network. In addition, each individual’s VKI data is calculated and placed into one of the four groups of data. The data is created to learn the model. In the deep nerve network, the data is divided into 80% training as 20% test. Training and testing processes have been carried out after the determination of the hyper parameters of the network. The test resulted in 93.2 percent accuracy. As a result, the recommended classification model with the individual's VKI determined obesity status is verified by the classification of foot pressure data. It is also observed that LSTM-based deep nerve networks have high success performance in classification processes.

Keywords: