RCDMS - Robust Car Detection & Modication System

See It. Segment It. Swap It. — Smart Car Part AI at Your Fingertips.

Comment 1

Robust Car Detection & Modification System - RCDMS

AI-Powered Car Part Detection & Modification, Made Effortless.

RCDMS (Robust Car Detection & Modification System) is a smart web-based SaaS platform that uses advanced AI models like YOLO and SAM to detect and segment car parts from real images. Users can visualize modifications—such as headlights, bumpers & side mirrors etc—in real time, overlaid on their actual vehicle. This not only improves customer satisfaction but also reduces errors and streamlines the entire customization workflow, making the process faster, smarter, and more cost-efficient.

Overview

This project is an end-to-end web platform that leverages state-of-the-art AI to detect, segment, extract, and visually modify car parts in images. Designed for automotive professionals, insurance companies, car enthusiasts, and researchers, it provides a seamless workflow from image upload to part replacement—all in your browser.

Features

  • Advanced Detection: YOLO-based car part detection with high accuracy.
  • Interactive Segmentation: Visualize and select car parts with color-coded masks.
  • Part Extraction & Removal: Instantly extract or mask out detected parts.
  • Visual Replacement: Upload and blend new parts into the original image.
  • Modern UI: Beautiful, responsive, and intuitive interface inspired by industry leaders.
  • Secure & Private: All processing is local to your session; your data is never shared.

How It Works

  1. Landing Page: Welcomes users with a modern, animated design and a clear call to action.
  2. Step-by-Step Workflow:
    • Upload Image: Start with any car image.
    • Detect Parts: AI model finds and labels car parts with visible bounding boxes.
    • Segmentation: Color-coded masks highlight each part for easy selection.
    • Extract/Remove: Instantly extract or mask out any detected part.
    • Replace: Upload a new part image and blend it into the original.
  3. Results: Download or share your modified image.

Tech Stack

  • Frontend: HTML5, CSS3, Bootstrap 5, JavaScript (vanilla)
  • Backend: Python 3, Flask
  • AI Models:
    • YOLO (for detection)
    • SAM (Segment Anything Model)
    • PyTorch, OpenCV, NumPy
  • Other:
    • Responsive design
    • SVG pattern backgrounds

Setup Instructions

  1. Clone the Repository bash git clone <your-repo-url> cd "RCDMS"
  2. (Optional) Create a Virtual Environment bash python -m venv ml # On Windows: ml\Scripts\activate # On Mac/Linux: source ml/bin/activate

  3. Download Segment Anything Model Link Here

  4. Install Dependencies

    pip install -r requirements.txt

    If you don't have a requirements.txt, install at least:

    pip install flask ultralytics segment-anything opencv-python numpy

  5. Run the App

    python app.py

  6. Open in Browser

Usage Guide

  1. Landing Page: Click "Get Started".
  2. Step 1: Upload a car image.
  3. Step 2: Click "Detect Parts" to see bounding boxes and labels.
  4. Step 3: Click "Apply Segmentation" for color-coded masks.
  5. Step 4: Select and extract/remove parts as needed.
  6. Step 5: Upload a replacement part and blend it in.
  7. Download or share your results!

Folder Structure

Machine Learning Hackathon/
├── app.py
├── requirements.txt
├── static/
│   ├── js/
│   │   └── main.js
│   ├── uploads/
│   
├── templates/
│   ├── landing.html
│   └── index.html
├── car_parts_detector.pt
├── sam_vit_h_4b8939.pth
├── yolo11s-seg.pt
└── ...

Special Thanks: Open-source contributors and the hackathon organizers.

Built With

Share this project:

Updates