VIDEO TRANSCRIPTION

  • Unique Paper ID: 171601
  • Volume: 11
  • Issue: 8
  • PageNo: 933-936
  • Abstract:
  • This project presents a web application for automated keyword extraction and analysis from video and audio content, including YouTube videos. Leveraging the Whisper speech recognition model and advanced natural language processing techniques, the application aims to provide efficient and accurate keyword extraction capabilities for various applications. The system processes uploaded video files or YouTube URLs, extracting audio content and transcribing it using Whisper. Subsequently, TF-IDF or CountVectorizer algorithms are employed to identify significant keywords from the transcribed text. For YouTube videos, keyword accuracy is evaluated against available transcripts, while for uploaded files, user-provided expected keywords can be used for comparison. Performance metrics, including precision, recall, F1-score, and coverage score, are calculated to assess keyword extraction effectiveness. The developed web application, built using the Flask framework, offers a user-friendly interface for seamless interaction and facilitates access to extracted keywords, full transcripts, and performance metrics. This tool holds potential applications in content analysis, information retrieval, video indexing, and automated metadata generation. The project demonstrates the feasibility of combining speech recognition and natural language processing for robust keyword extraction from multimedia content, offering valuable insights and efficiency improvements for content management and analysis workflows.

Cite This Article

  • ISSN: 2349-6002
  • Volume: 11
  • Issue: 8
  • PageNo: 933-936

VIDEO TRANSCRIPTION

Related Articles