Enhanced human activity recognition using 2DCNN, BiLSTM, and soft attention mechanism

Key Points

• Enhanced recognition accuracy is achieved through the integration of 2DCNN and BiLSTM.
• The method leverages a soft attention mechanism to effectively capture temporal dependencies in data.
• Key features are extracted from video sequences using 2D convolutional neural networks.
• This approach supports advancements in automated human activity recognition in diverse applications.