AI-Powered System for Detecting Offensive Meme Texts on Social Media

In the current digital environment, Sentiment analysis has become an essential tool for understanding human emotions and perspectives. Memes are a popular internet expression tool that frequently expresses opinions in an eye-catching manner. However, worries regarding the transmission of inappropriate or dangerous information have been raised by the emergence of multimodal content, particularly memes that combine text and graphics. In order This work offers a fresh solution to this problem for categorising objectionable memes in multimodal datasets using deep learning techniques. By closely analyzing both textual and visual components, the primary objective is to accurately detect and categorize inappropriate content. In this study, we use deep learning techniques to propose a novel approach for sentiment analysis of meme photos for text and image categorization using Convolutional Neural Networks in conjunction with Optical Character Recognition (OCR) technology. Our technology is based on extracting textual information from meme photos using optical character recognition (OCR), which enables sentiment analysis of these textual components. Deep learning techniques are included into this procedure, To classify sentiments extracted from the detected text, natural language processing techniques are used in conjunction with the (VADER). By employing this novel approach, we want to enhance visual content's analytical potential by utilising a pretrained Convolutional Neural Network (CNN) model. The suggested method yields promising results, demonstrating the possibility of integrating pretrained CNNs, To better understand online sentiment patterns and interpret the complex emotions present in meme culture, VADER sentiment analysis and OCR are being used.