INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
With the exponential growth of digital video content and the easy availability of online platforms, users are faced with the daunting task of sifting through vast amounts of video data to find relevant information. This project addresses this challenge by introducing a video summarization system that offers users an efficient solution for extracting important content from YouTube videos. The system uses state-of-the-art technologies, including the Hugging Face ASR (Automatic Speech Recognition) algorithm for speech-to-text conversion and a transformer-based model for text summarization. The process involves getting the URL of a YouTube video from the user, extracting the audio of the video, and generating the corresponding transcript using the PyTube library and the YouTube Transcript API. In parallel, the video audio is processed using the ASR algorithm to obtain additional text. The combined texts are then preprocessed and the transformer model is applied to produce a concise summary. To improve the user experience, summarized text can be converted to speech using a text-to-speech algorithm. The system includes various tools and libraries such as PyTorch, NLTK (Natural Language Toolkit) and Librosa for efficient audio processing, text analysis and feature extraction. The user interface provides a user-friendly interface with a text field for entering the URL of a YouTube video, a "Summary" button to start the summarization process, and an audio play button to listen to the audio version of the summary. The project aims to change the way users browse and consume video content by providing a time- efficient and comprehensive approach to extracting valuable information from videos. The result of the work is a ROUGE score of 0.95 that signifies an extremely good quality with a high degree of similarity between the original video and the summarized content.
"Video Summarization using NLP", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.8, Issue 6, page no.462 - 465, June-2023, Available :http://www.ijsdr.org/papers/IJSDR2306069.pdf
Downloads:
000337351
Publication Details:
Published Paper ID: IJSDR2306069
Registration ID:207012
Published In: Volume 8 Issue 6, June-2023
DOI (Digital Object Identifier):
Page No: 462 - 465
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn