Social Media Censoring Using of Visual Sentimental Analyzer
Y Swathi, K Krubasagare, Praveen. S, Ronald Ryan · 2024
This research paper proposes a solution to enhance the efficiency of social media platforms in classifying, flagging, and removing potentially vulgar or offensive posts. The aim is to reduce the time taken for such actions by implementing real-time image classification during the uploading process. This dynamic approach is facilitated by leveraging the power of transfer learning through the implementation of Inception v3, a highly accurate pre- trained image recognition model with extensive dataset support. Additionally, the paper presents the development of a user- friendly social media application akin to popular platforms like Instagram and Facebook. The application enables users to upload images accompanied by captions, browse through the posts of their followers, express appreciation through likes, and engage in discussions via comments. Furthermore, users have the ability to search for specific profiles, among other features commonly found in leading social media apps. The primary objective of this paper is to identify and address the limitations of existing image classification systems by introducing a novel model. The proposed model serves as the foundation for building an improved social media application, thus contributing to the overall enhancement of content moderation processes.