Image Augmentation Library

A Python library for image augmentation using OpenCV and YOLO. This library provides a variety of functions to manipulate images, such as adjusting HSV values, brightness, contrast, applying blur, rotating/scaling, and changing backgrounds using object detection and segmentation.

Features

• HSV Adjustment: Modify hue, saturation, and value of images.
• Brightness Adjustment: Add or subtract brightness from images.
• Contrast Adjustment: Increase or decrease image contrast using gamma correction.
• Blur Filter: Apply Gaussian blur with adaptive kernel size.
• Rotate and Scale: Rotate and scale images with adjustable parameters.
• Background Change (Simple): Use YOLOv8 bounding box detection to replace the background of detected objects (e.g., persons).
• Background Change (Segmentation): Use YOLOv8 segmentation to precisely replace the background of detected objects.
• Batch Processing: Apply augmentations to all images in a directory and save the results.

Installation

You can install the image_augmentation library via pip:

pip install image_augmentation

Prerequisites

Ensure you have the following dependencies installed:

• Python 3.6 or higher
• numpy>=1.19.0
• opencv-python>=4.5.0
• ultralytics>=8.0.0 (for YOLOv8 models)

You can install all dependencies automatically when you install the package via pip.

Usage

The library provides a main function augment that can apply various operations to a directory of images. Below are some examples of how to use the library:

Example 1: Increase Brightness

Increase the brightness of all images in a directory by adding a value of 100:

from image_augmentation import augment

augment(
    img_load_path="/path/to/input",
    img_save_path="/path/to/output",
    oper="brightness",
    operation="add",
    value=100
)

Available Operations

The augment function supports the following operations:

Operation	Description	Required Parameters
hsv	Adjust HSV values (hue, saturation, value)	hue, saturation, value
brightness	Adjust brightness (add or subtract)	operation ("add" or "subtract"), value
contrast	Adjust contrast using gamma correction	gamma (float)
blur	Apply Gaussian blur with adaptive kernel size	threshold, high_th, low_th
rotate_and_scale	Rotate and scale images	rotation_amount_degree, scale_factor
change_background_simple	Replace background using YOLO bounding boxes	background_img (path or numpy array)
change_background_segment	Replace background using YOLO segmentation	background_img (path or numpy array)

Project Structure

image_augmentation/
│
├── image_augmentation/       # Main package directory
│   ├── __init__.py           # Package initialization
│   ├── functions.py          # Core augmentation functions
│   ├── augment.py            # Main augment function for batch processing
│
├── tests/                    # Test directory (optional)
│   ├── test.py  # Unit tests
│
├── setup.py                  # Setup script for installation
├── README.md                 # Project documentation

Acknowledgments

• This library uses OpenCV for image processing.
• Ultralytics YOLOv8 is used for object detection and segmentation.
• Thanks to the open-source community for providing tools and inspiration.