Authors: Sai Kumar S L, Ramu B T, Mallikarjun Heroor, B M Shree Lakshmi, Dr. Mydhili Nair
Abstract: For exponential growth of user-generated content (UGC) on video-sharing platforms necessitates the development of highly efficient and scalable automatic content moderation and spam detection algorithms. Traditional manual review techniques are overwhelmed by the sheer volume and real-time nature of video uploads, which leads to unequal enforcement, moderator fatigue, and prolonged exposure to harmful content. This work offers a unique, multi-modal Video Content Moderation and Spam Detection tool that applies artificial intelligence and machine learning to handle these problems. To detect violent, sexually explicit, and policy-violating pictures, the system incorporates sophisticated Computer Vision (CV) techniques, such as frame-by-frame analysis, object detection, and visual hashing in order to identify hate speech, harassment, fraudulent schemes, and spam indications (such as harmful URLs, repetitive content, and behavioural anomalies), Additionally, it analyses video titles, descriptions, and comments.
DOI: http://doi.org/10.5281/zenodo.20758833
