https://github.com/lucasjacks0n/transummary

https://github.com/lucasjacks0n/transummary

Last synced: about 2 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/lucasjacks0n/transummary
• Owner: lucasjacks0n
• Created: 2023-11-15T20:11:33.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2023-11-22T21:20:49.000Z (about 1 year ago)
• Last Synced: 2024-04-24T08:54:11.992Z (8 months ago)
• Language: Python
• Size: 654 KB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# TranSummary

## Description

TranSummary is a web application designed to enhance the accessibility and comprehension of YouTube videos. It features a React frontend and a Flask API backend. Users can input a YouTube video URL into the frontend, which then interacts with the backend to perform a series of operations. The API downloads the video, transcribes it using Whisper diarization, extracts faces, and finally, summarizes the content into well-defined chapters with titles, leveraging the OpenAI API.

## Features

- **React Frontend**: An interface for submitting YouTube video URLs and viewing chapter summaries.

- **Flask API Backend**: Handles video processing and transcription.

- **Video Transcription**: Utilizes Whisper AI for precise transcription with timestamps.

- **Unique Face Detection**: Uses face_detection library to extract boxes and face features and DBScan to cluster for uniqueness

- **Content Summarization**: Implements OpenAI API to create concise, titled chapters for easy navigation and understanding of the video content.

![screenshot](transummary.jpg)

## References

- https://github.com/MahmoudAshraf97/whisper-diarization

- https://github.com/cppxaxa/FaceRecognitionPipeline_GeeksForGeeks