Emosense – Emotion Recognition using Images and Audio

Emosense – Emotion Recognition using Images and Audio

Prof. G. H. Wani1, Prathamesh Chaudhary2, Anurag Kotwal3, Raghvendra Kanakdande4, Shlok Das5

1,2,3,4,5Department of Artificial Intelligence and Data Science, AISSMS Institute of Information Technology, Pune, Maharashtra, India - 411001

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - With the combination of audio and visual analysis, emotion recognition—a crucial aspect of human-computer interaction—has advanced significantly. This multidisciplinary method uses image processing and machine learning methods to extract emotional states from multimodal data. Pitch, tone, and speech patterns are among the characteristics that are extracted in audio analysis in order to infer emotions like happiness, rage, or melancholy. Simultaneously, image analysis uses facial expression recognition to interpret facial motions and micro expressions as indicators of emotion. When these modalities are combined, emotion detection systems—which are useful in domains like healthcare, human-computer interfaces, and sentiment analysis for better human-AI interaction—become more accurate, robust, and applicable in real time.

Key Words:  Emotion recognition, Audio and Image, Multimodal approach, Machine learning, Deep learning, Data preprocessing, Feature extraction, Real-time emotion recognition, Human-computer interaction.