MULTIMODEL EMOTION DETECTION IN VIDEOS USING PRE-TRAINED LLMS

Abstract

This paper presents a comprehensive system designed to detect emotions in videos using pre-trained Large Language Models (LLMs). The system integrates video processing and natural language processing techniques to analyze both visual and audio data, providing a robust solution for emotion detection. This paper details the system architecture, implementation, testing methodologies, and the tangible benefits observed during initial deployments.