Skip to content

mrfree96/emotionAI

Repository files navigation

EmotionAI: Emotion Analysis from Video

Overview

EmotionAI leverages the power of machine learning to analyze emotions from video content with high accuracy. It integrates three sophisticated models, each dedicated to interpreting emotional cues from different aspects of the video:

  • Facial Expressions: Identifies emotions through the analysis of subtle facial movements.
  • Language Model for Textual Transcription: Employs a language model to predict emotions from the transcription(Speech to text by OpenAI's Whisper library).
  • Audio Signals: Analyzes background noises and vocal sounds to add an additional layer of emotional context.

This diverse approach guarantees a thorough study. A complex understanding of emotions is made possible by the integration and weighted contributions of each model's output.

App Link

Discover the depth of emotional analysis with EmotionAI: EmotionAI Streamlit App

Getting Started

Engage with EmotionAI by simply uploading a video. Our algorithms meticulously analyze each frame, assessing facial expressions, utilizing OpenAI's Whisper library to transcribe speech, and employing a language model to predict emotions from the transcription. Additionally, audio signals are evaluated to deliver a comprehensive emotional analysis.

Features

  • In-depth Emotion Analysis: Combines facial, textual, and auditory cues to provide a comprehensive understanding of emotional states.
  • Intuitive Interface: A user-friendly platform that simplifies the process of uploading and analyzing video content.
  • Advanced Insights: Delivers profound insights into the emotional undertones of videos, powered by cutting-edge analytics.

Contributions

Contributions are warmly welcomed. Whether you're enhancing the models, refining the code, or offering feedback on usability, your input is invaluable.

License

EmotionAI is proudly open source, available under the MIT License.

About

Emotion Analysis from Video: A Multi-Model Machine Learning Approach

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages