Paper Contents
Abstract
Ensuring the safety, quality, and authenticity of medicines is a critical requirement in modern healthcare systems. Simultaneously, the early detection of life-threatening diseases such as brain tumors and chest infections demands reliable and accurate diagnostic models. This paper presents MediSure AI, a multimodal artificial intelligence framework that integrates machine learning, deep learning, and retrieval-augmented generation (RAG) to support pharmaceutical safety evaluation and medical image analysis. The system employs a Random Forest classifier for structured medicine-parameter assessment, VGG16 for tablet image classification, and ResNet50 for disease prediction using MRI and chest-X-ray datasets. Additionally, a KMeansPCA clustering module enables pattern discovery within pharmaceutical records, while a RAG-based medical chatbot powered by Gemini enhances user interaction through context-aware query resolution. Experimental results demonstrate high accuracy across all modules, including 98% for medicine safety prediction and over 90% for brain tumor classification. The system is implemented through a lightweight interactive interface and optimized preprocessing pipelines, ensuring practical usability for pharmacists, clinicians, and academic users. Overall, MediSure AI highlights the potential of multimodal AI systems to deliver efficient, scalable, and accessible healthcare intelligence.Index Terms Medicine safety prediction, deep learning, medical imaging, Random Forest, VGG16, ResNet50, RAG chatbot, healthcare AI.
Copyright
Copyright © 2025 Preet Sharma, Sushant Shekhar, Ashit Vijay, Sahil Yadav. This is an open access article distributed under the Creative Commons Attribution License.
