SentiOCR — Advanced Handwritten Sentiment Analysis

Live Site: https://frontend-37g8xavhf-hackersmkgs-projects.vercel.app

SentiOCR is a high-performance, full-stack AI platform designed to bridge the gap between physical handwriting and digital emotional intelligence. It leverages OpenCV-enhanced OCR to extract text from handwritten images and applies a weighted fuzzy-matching algorithm against a dynamic SQLite sentiment dictionary to classify emotional tone.

🚀 Features

• OCR Engine: OpenCV-preprocessed Tesseract OCR for high-accuracy text extraction.
• Sentiment Logic: Multi-weighted sentiment scoring with fuzzy word matching.
• Dynamic Dictionary: Manage positive, negative, and neutral keywords via the UI.
• Modern UI: Responsive React + TypeScript frontend with drag-and-drop support.

🛠️ Tech Stack

• Frontend: React, TypeScript, Vite, Tailwind CSS, Lucide Icons.
• Backend: FastAPI, SQLAlchemy (SQLite), Pytesseract, OpenCV.
• Deployment: Vercel (Frontend) & Render (Backend).

📦 Installation & Setup

1. Backend (Python)

cd backend
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn app:app --reload

2. Frontend (React)

cd frontend
npm install
npm run dev

🌐 Deployment Instructions

Frontend (Vercel)

1. Connect your GitHub repository to Vercel.
2. Set the Root Directory to frontend.
3. Add Environment Variable: VITE_API_URL = (Your Render Backend URL).
4. Deploy.

Backend (Render)

1. Create a Web Service on Render.
2. Set the Root Directory to backend.
3. Environment: Docker (This is critical to ensure Tesseract OCR is installed).
4. Render will automatically detect the Dockerfile and build the image.

📜 License

MIT