Transplant Track: Video Content Summarization with Whisper and GPT Transformers
  • Author(s): Kamsani Rohith Reddy; Vasam Akhilesh; Kundanapally Rakesh; Halavath Balaji
  • Paper ID: 1716256
  • Page: 1220-1225
  • Published Date: 14-04-2026
  • Published In: Iconic Research And Engineering Journals
  • Publisher: IRE Journals
  • e-ISSN: 2456-8880
  • Volume/Issue: Volume 9 Issue 10 April-2026
Abstract

The scalding development of the procedure of creation of digital material has led to bursting of the demand of the short form of the video contents of the reels, shorts and clips that may keep the audience entertained during a finite duration. The long form videos may however be time and labour-intensive since the content creators will have to choose interesting highlights manually. The paper introduces a Video-to-Reel Conversion Platform which is an intelligent platform to convert long videos to small and entertaining short reels. It is referred to as a speech recognizer, sentiment analysis, multimedia and artificial intelligence that is able to extract a valuable content of video information of a long length. The interactive interface of the site may be created with the assistance of the Streamlit and allows individuals to visit the site with the help of offering videos or links on YouTube to be processed. The videos uploaded are stored in a PostgreSQL database such that they are safe and easily accessed. FFmpeg is used to remove the audio track and Whisper, a speech recognition model that uses the approaches of the deep-learning model, is used to produce relevant text transcripts. The analysis is further done with the help of GPT and TextBlob to find out the intensity of emotion, relatability of the subject and interest of the reader. Highlight selection algorithm ranks most interesting contents of the videos according to the following characteristics of transcription sentiment polarity, speech pacing and keyword density which rank the videos. The selected parts will be automatically cut and submitted to FFmpeg and MoviePy to create quality reels in small format. The last reels are made available to the users via the platform by the option of downloading or sharing via the digital platforms. Using the analysis of the experiment, we may observe that the proposed system can contribute to the significant drop in the number of manual laborers to be employed to create the short video highlights and be rather topical and interesting. The automated pipeline is offering a valuable and scalable solution to content creators, marketers, and media employees who would prefer to use long-form video content to produce entertaining short-form media.

Keywords

Video Highlight Detection, Automated Reel Generation, Speech Recognition, Whisper Model, Sentiment Analysis, FFmpeg, Moviepy, Artificial intelligence, Streamlit, Video content processing.

Citations

IRE Journals:
Kamsani Rohith Reddy, Vasam Akhilesh, Kundanapally Rakesh, Halavath Balaji "Transplant Track: Video Content Summarization with Whisper and GPT Transformers" Iconic Research And Engineering Journals Volume 9 Issue 10 2026 Page 1220-1225 https://doi.org/10.64388/IREV9I10-1716256

IEEE:
Kamsani Rohith Reddy, Vasam Akhilesh, Kundanapally Rakesh, Halavath Balaji "Transplant Track: Video Content Summarization with Whisper and GPT Transformers" Iconic Research And Engineering Journals, 9(10) https://doi.org/10.64388/IREV9I10-1716256