A Novel Approach for Inappropriate Content Detection and Classification of Youtube Videos using Deep Learning

The rapid expansion of YouTube videos has attracted billions of viewers, the majority of whom are young. Nevertheless, fraudulent uploaders also take advantage of this site to disseminate disturbing videos to everyone, especially kids, like sharing indecent content through cartoons and movies but within all of this content comes unsuitable content that can be dangerous to viewers, particularly for young viewers. The appearance of such unsuitable content poses serious concerns, particularly for young viewers. Thus, it is highly recommended that social platforms incorporate real-time video content filtering. This study has proposed a new method which is an innovative deep learning architecture for identifying and categorizing objectionable video content is presented in this proposal. In particular, it extracts video descriptors by using Generative Adversarial Networks and Our method improves inappropriate video content detection and classification for child safety by better capturing the contextual information of video descriptors, as confirmed by comparisons with existing methodologies. Besides technological innovation, our principal objective is to contribute to the development of safer online environments. We hope to strengthen moderation of content by leveraging deep learning, resulting in a safer online landscape for people all around the world.