Multi-Modal Summarization for Digital Data

Abstract

ABSTRACT
Text summarization aims to abbreviate input text into a compact form. Automatic data summarization is one of the field of data mining. The increase in expansion of data communication over the net makes it crucial to create Multi-Modal Summarization (MMS) from synchronous data (text, audio, image and video). The MMS method combining the techniques of Natural Language Processing (NLP), speech processing, computer vision, and Advanced Encryption Standard (AES) encryption used to traverse the abundant information present in multi-modal data, to improve quality as well security. The main scheme is to bridge and lessen the allowable gaps between multi-modal data. Video is consisting of audio and visual (image). Speech transcriptions used for audio. For observable information, the joint representations of image and text are studied using an artificial neural network for visual information. Finally, all the multi-modal features are examined to generate a textual summary by increasing the salience, readability. The summary generated by text, audio, image, or video is encrypted using the AES data encryption method (to make it secure) and stock on the server, from where the user can retrieve it when required by supplying the decryption key.

Keywords: AES, encryption, decryption, summarization, multimedia, multi-modal, natural language processing

Cite this Article: Malan D. Sale, Archana Mahajan, Akshay Nale, Yogesh Shelar, Asawari Melgire. Multi-Modal Summarization for Digital Data. International Journal of Embedded Systems and Emerging Technologies. 2020; 6(1): 28–34p.