Deep Learning Methodologies For Human Activity Recognition

An improved BiLSTM model for daily activity recognition

“…Alaghbari et al [10] proposed a Deep Neural Networks (DNN) daily activity recognition model to monitor the elderly in their daily lives, such as eating, sleeping or taking medicine. Alhumayyani et al [11] proposed three main deep learning methods, which are based on LSTM, BiLSTM and Gate Recurrent Unit (GRU), to recognize daily activity. Arifoglu et al [12] also used Recursive Neural Network (RNN), Vanilla Recursive Neural Network (VRNN), and LSTM for the identification of abnormal daily activities, respectively.…”