Copyright © 2025 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{177658,
        author = {Tejal Sudhakarrao Mohod and Dr. Ashish A. Bardekar},
        title = {A Prompt Based Modal for Hateful Meme Classification},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {12},
        pages = {1383-1388},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=177658},
        abstract = {Hateful memes have emerged as a potent vehicle for spreading toxic content on social media by combining seemingly benign visuals wih offensive or harmful text. Detecting such content poses a significant challenge due to its multimodal nature, where the hateful intent may lie in the interplay between image and text. In this paper, we propose a prompt-based approach for multimodal hateful meme classification using a combination of Optical Character Recognition (OCR), vision-language captioning, and transformer-based textual analysis. Our system extracts embedded text using EasyOCR, generates contextual image captions via the BLIP model, and classifies the combined text using a fine-tuned RoBERTa model. In addition to hatefulness classification, we extend our framework to perform emotion detection on the meme content, providing deeper insights into the emotional tone. Furthermore, we introduce an automatic PDF report generation module that consolidates the analysis results for practical use cases. Experimental results demonstrate competitive performance on benchmark datasets, showcasing the effectiveness of integrating visual and linguistic cues. This work highlights the potential of prompt-based transformer models in addressing the complex task of multimodal hate speech detection while offering a comprehensive analysis pipeline.},
        keywords = {BLIP, EasyOCR, hateful meme detection, multimodal learning, RoBERTa, emotion detection, vision-language models, report generation.},
        month = {May},
        }