Abstract:
"In the realm of mental health monitoring and emotion awareness, the integration of multimodal emotion recognition and emoji generation systems emerges as a promising avenue. This research endeavours to develop a comprehensive framework that harnesses the power of various sensory modalities, including facial expressions, speech patterns, and physiological signals, to accurately detect and analyse individuals' emotional states. The proposed system not only focuses on recognizing emotions but also seeks to bridge the communication gap by introducing a novel emoji generation component. This innovative feature aims to enhance user engagement and self-expression, providing a dynamic and user-friendly interface for individuals to convey their emotional experiences.
The project's significance lies in its potential to revolutionize mental health monitoring, offering a non-intrusive and continuous assessment tool. By leveraging machine learning algorithms and deep neural networks, the system aims to achieve high accuracy in emotion recognition across diverse contexts. Furthermore, the emoji generation module aims to empower users by enabling them to articulate their emotions visually, fostering a deeper understanding of one's emotional well-being.
According to the preliminary test findings, the face emotion recognition model exhibits an accuracy of 0.8373, signifying that the model accurately classifies 83.73% of instances. In the case of the speech emotion recognition model, the accuracy is recorded at 97%, representing the percentage of correct predictions out of all instances. The precision for the chosen classes is 0.97, denoting the ratio of true positive predictions to all positive predictions."
