VanCorp Holdings

AI Virtual Try-On System

A Modular, Scalable, Production-Ready Virtual Try-On System powered by Deep Learning, designed to enable users to try clothes virtually using just an image. Powered by DL & vision.

      +----------------+       +------------------+        +-----------------+
      |   User Upload  | --->  |  Pose Estimation  |  --->  | Human Parsing   |
      +----------------+       +------------------+        +-----------------+
               |                         |                          |
               |                         v                          v
               |            +-----------------------+     +------------------+
               |            | Cloth Warping (TPS)   | <-- | Clothing Upload  |
               |            +-----------------------+     +------------------+
               |                         |
               v                         v
      +------------------------------------------+
      |        Fusion & Rendering Module         |
      +------------------------------------------+
                          |
                          v
           +-------------------------------+
           |  Output Final Try-On Result   |
           +-------------------------------+

⚙️ Tech Stack

Component Technology Used Framework FastAPI (Python) Pose Estimation MediaPipe Pose / HRNet / OpenPose Cloth Segmentation OpenCV (baseline), CIHP Parsing, DeepLabV3+ Warping Engine Thin Plate Spline (TPS), CP-VTON, VITON-HD Fusion Generator TryOnGAN, SPADE, Diffusers (Stable Diffusion) Image I/O OpenCV, NumPy Real-time Support ONNX, TensorFlow Lite, MediaPipe

🚀 Features

✅ Upload body and cloth images ✅ Estimate body landmarks ✅ Segment and isolate clothing regions ✅ Wrap clothes using keypoints (TPS) ✅ Fuse cloth realistically using GAN or Diffusion ✅ Save and serve final rendered try-on result ✅ Easily extendable and modular

Project Structure

virtual try on/ ├── app/ │ ├── main.py │ ├── routes/ │ │ ├── upload.py │ │ └── tryon.py │ ├── ml/ │ │ ├── pose_estimator.py │ │ ├── cloth_segmentation.py │ │ ├── cloth_warping.py │ │ ├── fusion.py │ │ └── tryon_pipeline.py │ └── utils/ │ └── image_io.py ├── static/ │ ├── uploads/ │ │ ├── users/ │ │ └── cloths/ │ └── outputs/ ├── models/ │ ├── cihp/ │ ├── viton/ │ └── tryongan/ ├── requirements.txt └── README.md

📦 Setup Instructions

1️⃣ Clone the Repository

git clone https://github.com/your-username/google-tryon.git cd "your file name"

2️⃣ Create Environment

python -m venv venv venv\Scripts\activate # Windows #source venv/bin/activate # Linux/Mac

3️⃣ Install Dependencies

pip install -r requirements.txt ⚠️ Make sure you use Python 3.10 for MediaPipe & DL compatibility.

4️⃣ Start the FastAPI Server

uvicorn app.main:app --reload

Access Swagger UI at: http://localhost:8000/docs

🖥️ Minimum System Requirements

CPU: Intel i7/i9 (8+ cores)

GPU: NVIDIA RTX 3060/4060 or higher

RAM: 16GB+

Disk: 10GB+ free (for model weights)

Python: 3.10.x

Contributors

~ Satyabrat Sahu