Techniques for Meeting Summarization: An Analysis and Suggestions for Improvement

Authors

  • Viveksheel Yadav, Faraz Ahmad, Ashuvendra Singh

DOI:

https://doi.org/10.17762/msea.v71i4.688

Abstract

The purpose of this research is to investigate the several approaches to summarizing meeting minutes. In addition to this, it proposes a hybridized strategy for the summarizing of meetings, which brings together abstractive and extractive methods. Audio extraction and text detection from the screen are the two methods that may be used in order to successfully extract text from meetings. After the necessary texts have been collected, a meeting recap is generated from them. Abstractive text summarizing, on the other hand, is more concerned with presenting a logical overview of the material that has been presented, as opposed to extractive text summarization, which consists of stringing together the most important lines from various paragraphs. The bulk of recent research on abstract text summarization has been built on recurrent neural networks; nevertheless, RNN-based algorithms do not perform well when dealing with lengthy sequences. [Citation needed] [Citation needed] We recommend combining these two approaches to the summarization of meetings in a sequential fashion in order to provide a more effective overall summary of the meeting. Our hybridized text summarizer model is going to be applied to the conference video and audio once it has been converted into text documents as part of the recommended research endeavor. The final product would be a condensed text document that the different parties involved, whether or not they were physically present at the conference, would be able to use as a quick reference that gets to the point.

Downloads

Published

2022-09-02

How to Cite

Viveksheel Yadav, Faraz Ahmad, Ashuvendra Singh. (2022). Techniques for Meeting Summarization: An Analysis and Suggestions for Improvement. Mathematical Statistician and Engineering Applications, 71(4), 1634–1645. https://doi.org/10.17762/msea.v71i4.688

Issue

Section

Articles