This research introduces a lightweight, multimodal model for Medical Visual Question Answering that combines BiomedCLIP for image processing and LLaMA-3 for text analysis.

• Achieves state-of-the-art performance while using fewer computational resources
• Effectively handles diverse medical imaging modalities (X-rays, CT scans, etc.)
• Optimizes clinical decision support through natural language querying of medical images
• Demonstrates how specialized AI architectures can balance performance with efficiency

Why It Matters: This approach makes advanced medical image analysis more accessible for clinical environments with limited computational resources, potentially accelerating adoption of AI-assisted diagnostics and improving patient care.

A Lightweight Large Vision-language Model for Multimodal Medical Images