Ensemble Learning Using Pressure Sensor for Gait Recognition

This paper proposes an ensemble deep learning model that can identify each person using pressure sensor data acquired from smart insoles. The ensemble model consists of CNN, LSTM, and self-attention to effectively learn the relationship between gait data. The model was trained using a triplet loss to map each data to a better embedding vector of latent space. The experimental results showed that the accuracy was improved in recognizing people by using the proposed ensemble learning while reducing the size of the model by half.