Sports Activity Recognition with Deep Learning Models and Accelerometers
Hoang-Dieu Vu, Quang-Tu Pham, Duc-Nghia Tran, Hoang-Nam Le, Dinh-Dat Pham, Van-Toi Nguyen, To-Hieu Dao, Duc–Tan Tran · 2024
Sport activity classification has become increasingly vital in various domains, including scientific research and healthcare. Accurate classification of Sports activities holds significant implications for behavior analysis and medical diagnostics. This paper investigates the classification of Sports activities using deep learning models such as Multilayer Perceptron (MLP), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM) networks applied to accelerometer sensor data. The accelerometer data is collected from a wearable device affixed to the waist of individuals engaged in daily activities. The performance of the deep learning models is evaluated by training them for 100 epochs and assessing their accuracy and confusion matrices on a validation dataset. The primary aim of this research is to compare the effectiveness of different deep learning models in accurately classifying Sports activities based on accelerometer data.