Description of Invention
To address the need for effective emotion recognition, this project develops a Gait-Based Emotion Recognition system using an enhanced pre-trained InceptionV3 model. It analyzes individuals’ walking patterns from video inputs to classify emotions such as happy, sad, angry, and neutral. Gait serves as a non-intrusive, passive biometric indicator of emotional state. The InceptionV3 model, originally trained on large-scale image datasets, is adapted to extract high-level features and boost classification performance. Feasibility and accuracy are assessed through systematic training and testing. This work highlights the potential of the enhanced model in improving non-invasive emotion recognition systems.