Deep_Cleanse

Deep_Cleanse is a full-stack media denoising platform that uses deep learning models to analyze and reduce noise in images and audio.

The project combines:

• A React frontend for user interaction and visualization
• A Node.js/Express backend for authentication and session persistence
• A Python/Flask backend for image/audio analysis and denoising inference

Repository Structure

Deep_Cleanse/
├── frontend/   # React application (UI + client logic)
├── backend/    # Node.js API (auth, user data, saved sessions)
├── Python/     # Flask + TensorFlow denoising APIs
└── README.md

Features

• Image noise analysis and denoising
• Audio noise analysis and denoising
• Progressive denoising rounds with metrics
• Session saving and retrieval for image/audio workflows
• Email verification flow for signup/login

Tech Stack

• Frontend: React, Axios, Chart.js
• Node Backend: Express, Mongoose, Nodemailer, Helmet
• Python Backend: Flask, TensorFlow, OpenCV, Librosa, SoundFile, Pillow
• Database: MongoDB

Prerequisites

Install the following before running the project:

• Node.js 14+ (LTS recommended)
• Python 3.8+
• pip
• MongoDB (local or cloud instance)

Installation

1) Install frontend dependencies

cd frontend
npm install

2) Install Node backend dependencies

cd ../backend
npm install

3) Install Python dependencies

From the repository root:

pip install flask flask-cors tensorflow pillow numpy opencv-python librosa soundfile

Environment Configuration

Create a .env file inside backend/:

PORT=5000
MONGODB_URI=<your_mongodb_connection_string>
EMAIL_USER=<your_email_address>
EMAIL_PASS=<your_email_app_password>

EMAIL_USER and EMAIL_PASS are used by the verification email flow.

Running the Application

Recommended (start all services together)

cd frontend
npm start

This command starts:

• Frontend on http://localhost:3000
• Python backend on http://localhost:5001
• Node backend on http://localhost:5000

Run services individually

Frontend

cd frontend
npm start

Node backend

cd backend
npm start

Python backend

cd Python
python app.py

API Overview

Python denoising API (port 5001)

• POST /api/image/analyze - Analyze image noise
• POST /api/image/denoise - Denoise image by image_id
• GET /api/image/save/<image_id> - Download denoised image
• POST /api/audio/analyze - Analyze uploaded audio file
• POST /api/audio/denoise - Denoise audio by audio_id
• GET /api/audio/save/<audio_id> - Download denoised audio
• GET /api/audio/preview/<audio_id> - Audio preview payload

Node API (port 5000)

• POST /api/signup - Register user (verification required)
• POST /api/verify-email - Verify registration code
• POST /api/resend-verification - Send new verification code
• POST /api/login - User login
• GET /api/user/:email - Fetch user profile
• POST /api/image/save-session - Save/update image denoising session
• GET /api/image/sessions/:email - List image sessions
• GET /api/image/session/:id - Get one image session
• POST /api/audio/save-session - Save/update audio denoising session
• GET /api/audio/sessions/:email - List audio sessions
• GET /api/audio/session/:id - Get one audio session

Build and Test

From frontend/:

npm run build
npm test

Troubleshooting

• If react-scripts: not found, run npm install inside frontend/.
• If backend modules are missing (for example dotenv), run npm install inside backend/.
• If the Node backend fails to connect, verify MONGODB_URI in backend/.env.

Additional Documentation

A detailed project document is also available at:

• DeepCleanse_Documentation.docx