YOLO Vision Inspector 🔍

A web-based tool built with FastAPI and OpenCV to visualize YOLO (You Only Look Once) annotations directly in your browser. It supports both standard Bounding Boxes (object detection) and Instance Segmentation masks.

---

✨ Key Features

• FastAPI Backend
• Web Interface
• Dual Visualization Modes: Switch between Bounding Box and Instance Segmentation.
• Dockerized: Fully containerized environment for consistent deployment across any OS.

---

Bounding Box Mode

• Format: <class_id> <x_center> <y_center> <width> <height> (normalized).

Instance Segmentation Mode

• Format: <class_id> <x1> <y1> <x2> <y2> ... <xn> <yn> (normalized).

---

🚀 Getting Started

The easiest way to run the YOLO Vision Inspector is using Docker. This ensures that all dependencies (like OpenCV and system libraries) are configured correctly without polluting your local environment.

Prerequisites

• Docker Desktop installed and running.
• Git installed.

Installation & Deployment

1. Clone the Repository

   git clone [https://github.com/csenri/YOLO-Vision-Inspector.git](https://github.com/csenri/YOLO-Vision-Inspector.git)
   cd YOLO-Vision-Inspector

2. Build the Docker Image This command downloads the base Python image, installs system dependencies (libgl1, libglib2.0-0), and sets up the Python environment.

   docker build -t yolo-inspector .

3. Run the Container Map port 8000 of your host computer to port 8000 of the container.

   docker run -p 8000:8000 yolo-inspector

4. Access the Application Open your web browser and navigate to: 👉 http://localhost:8000

---