🩺 DermaAssist AI Agent

An AI-powered dermatology assistant that combines Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), Computer Vision, and Intelligent Triage to assist in preliminary skin disease analysis.

The system allows users to describe symptoms, upload skin images, receive AI-powered disease predictions, retrieve relevant medical knowledge, and obtain an intelligent triage recommendation.

---

✨ Key Features

• 🤖 Gemini 2.5 Flash powered medical reasoning
• 📚 Retrieval-Augmented Generation (RAG)
• 🔍 FAISS Vector Database
• 🖼 AI Skin Disease Image Classification
• 🚨 Intelligent Triage Agent
• 💬 Natural Language Symptom Analysis
• 🧠 LangChain-based Agent Workflow
• 🌐 Interactive Gradio Web Interface

---

📖 Project Overview

DermaAssist AI Agent is an intelligent dermatology assistant designed to provide preliminary skin disease analysis by combining Large Language Models (LLMs), Computer Vision, and Retrieval-Augmented Generation (RAG).

The application accepts both text-based symptom descriptions and skin lesion images. It analyzes user symptoms using AI agents, predicts possible skin diseases from uploaded images, retrieves relevant medical knowledge from a FAISS vector database, and generates personalized recommendations along with an intelligent triage assessment.

⚠️ Disclaimer: This project is intended for educational and research purposes only. It is not a substitute for professional medical diagnosis or treatment.

---

🩺 DermaAssist AI Agent

AI-powered dermatology assistant built using Gemini 2.5 Flash, LangChain, FAISS, RAG, Gradio, and Deep Learning Image Classification.

⚙️ Tech Stack

Category	Technologies
Programming Language	Python
Large Language Model	Gemini 2.5 Flash
Framework	LangChain
Vector Database	FAISS
Embedding Model	Sentence Transformers
Computer Vision	PyTorch, Torchvision
Machine Learning	Transformers
Frontend	Gradio
Document Processing	PyPDF
Environment	Python Virtual Environment

---

🏗️ System Architecture

                    User
                      │
        ┌─────────────┴─────────────┐
        │                           │
        ▼                           ▼
 Symptom Description          Skin Image
        │                           │
        ▼                           ▼
   Intake Agent             Image Classifier
        │                           │
        ▼                           ▼
 Recommendation Agent        Disease Prediction
      (RAG + FAISS)
        │
        ▼
    Gemini 2.5 Flash
        │
        ▼
    Triage Agent
        │
        ▼
 Final Medical Recommendation

📷 Screenshots

Home

Symptoms Analysis

Image Prediction

Final Recommendation

🌐 Live Demo

🚧 Deployment to free cloud hosting is currently limited by memory constraints.

Run locally using:

python app.py

This is better than linking to a failed deployment.

---

## 🌐 Deployment

> **Note**
>
> The application runs successfully in the local environment.
>
> Deployment to free cloud platforms (Render, Hugging Face Spaces free tier) is limited by memory constraints because the application loads multiple AI components including:
>
> - Gemini 2.5 Flash
> - LangChain
> - FAISS Vector Database
> - Sentence Transformers
> - PyTorch Image Classification Model
>
> The repository includes the complete source code, screenshots, and setup instructions for running the project locally.

## 🚀 Installation

```bash
git clone https://github.com/yourusername/DermaAssist-AI-Agent.git

cd DermaAssist-AI-Agent

pip install -r requirements.txt

python app.py

📌 Future Improvements

• Multi-disease classification
• PDF report generation
• Patient history tracking
• Cloud deployment
• Dermatologist appointment integration