YOLOv8 Object Detection, Cropping, and Classification

#################### PROJECT IS IN DEVELOPMENT ####################

This repository contains a set of Python scripts for detecting objects using YOLOv8, capturing images of unknown objects, cropping those images based on detected bounding boxes, and automatically classifying the cropped images using clustering.

Table of Contents

• Introduction
• Requirements
• Installation
• Usage
  • Object Detection and Capture
  • Cropping Images
  • Automatic Classification
• Directory Structure
• Contributing
• License

Introduction

The repository includes three main scripts:

1. object_detector.py: Detects objects in real-time using YOLOv8, captures images of unknown objects, and saves them along with metadata.
2. crop_images.py: Crops the captured images based on the bounding boxes stored in the metadata.
3. auto_classifier.py: Automatically classifies the cropped images using clustering and user input for labeling.

Requirements

• Python 3.6 or higher
• OpenCV
• NumPy
• scikit-learn
• tqdm
• ultralytics (for YOLOv8)

Installation

1. Clone the repository:

   git clone https://github.com/0xroyce/yolov8-dct.git
   cd yolov8-dct

2. Create a virtual environment and activate it:

   python3 -m venv venv
   source venv/bin/activate

3. Install the required packages:

   pip install -r requirements.txt

Usage

Object Detection and Capture

Run object_detector.py to start detecting objects and capturing images of unknown objects.

python object_detector.py

This script uses your webcam to detect objects in real-time. Known objects are highlighted in green, while unknown objects are highlighted in red and captured for further processing.

Cropping Images

Run crop_images.py to crop the captured images based on the bounding boxes stored in the metadata.

python crop_images.py

This script processes the images in the unknown_objects directory and saves the cropped images in the cropped_unknown_objects directory.

Automatic Classification

Run auto_classifier.py to classify the cropped images using clustering.

python auto_classifier.py

This script uses KMeans clustering to group similar images together. You will be prompted to label each cluster, and the images will be renamed accordingly.

Directory Structure

• unknown_objects/: Contains the images and metadata of unknown objects captured by object_detector.py.
• cropped_unknown_objects/: Contains the cropped images processed by crop_images.py.

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

This project is licensed under the MIT License. See the LICENSE file for details.