SUMMARIZE PODCAST TRANSCRIPTS

SUMMARIZE PODCAST TRANSCRIPTS

Harshitha M1, Deepika M N2 , Lamees Fathima3 ,Sachin K K4 , Spoorthi R5

1Harshitha M, ISE, Vidya Vikas Institute of Engineering & Technology Mysore

2Deepika M N, ISE, Vidya Vikas Institute of Engineering & Technology Mysore

3Lamees Fathima, ISE, Vidya Vikas Institute of Engineering & Technology Mysore

4Sachin K K , ISE, Vidya Vikas Institute of Engineering & Technology Mysore

5Spoorthi R, ISE, Vidya Vikas Institute of Engineering & Technology Mysore

Abstract - In the modern world, a significant amount of videos with information are uploaded every day. Selecting the appropriate video and comprehending the proper content. Despite the abundance of videos available, some of them will contain pointless material, and even if there is flawless content, it should still be needed of us. If the correct one is not located, all of your time and effort will be wasted trying to extract the useful information. We offer a novel concept that makes use of NLP. This gives consumers the ability to distinguish between pertinent and irrelevant information based on their needs by providing the video's key material in text description and abstractive summary. Additionally, our tests demonstrate that the joint model can provide high-quality, succinct, and understandable multi-line video summaries and descriptions in a human evaluation text extraction through processing and text summarization through BERT summarization.

Key Words: Video Summarization, Deep Neural Networks, Supervised Learning, Summarization Datasets, Evaluation Protocols.