Artistic Circle Effect Generator

A powerful image and video processing tool that transforms visual content into an artistic stippling effect using dynamically sized white circles on a black background. This application processes images, videos, and live webcam input to create unique artistic renderings where brightness values control the size of circles.

Features

• Multiple Input Types: Process images, videos, or webcam input
• High Performance: Optimized for CPU with batch processing and vectorized operations
• Resolution Control: Automatically resizes large videos to Full HD (1920x1080) while preserving aspect ratio
• Batch Processing: Accelerates video processing by handling multiple frames simultaneously
• Real-time Feedback: Shows processing speed and progress information
• Flexible Output: Choose to save results or just view the processing
• Command-line Interface: Easy to use with customizable parameters

Installation

Prerequisites

• Python 3.7 or higher
• OpenCV
• NumPy

Setup

1. Clone this repository:

git clone https://github.com/LeDat98/Circleartify
cd Circleartify

2. Install required packages:

pip install opencv-python numpy

Usage

The application provides a simple command-line interface with various options:

python artistic_effect.py [input] [options]

Basic Examples

Process an image:

python artistic_effect.py image.jpg

Process a video:

python artistic_effect.py video.mp4

Use webcam as input:

python artistic_effect.py webcam

Advanced Examples

Process an image with a larger window size and display the result:

python artistic_effect.py image.jpg --window-size 10 --show=True

Process a video with custom settings:

python artistic_effect.py video.mp4 --window-size 8 --batch-size 15 --output processed_video.mp4

Process webcam feed without saving:

python artistic_effect.py webcam --save=False

Available Options

Option	Short	Default	Description
--output	-o	Auto-generated	Output file path
--window-size	-w	5	Size of the processing window
--batch-size	-b	10	Number of frames to process at once (videos only)
--save		True	Whether to save the output
--show		False	Whether to display processing in real-time

How It Works

The algorithm divides the input image or video frame into small square regions (windows) of specified size. For each window:

1. It calculates the average brightness of pixels in that window
2. Based on this brightness value, it determines a circle radius (brighter regions get larger circles)
3. It draws a white circle at the center of each window with the calculated radius
4. The collection of these circles forms the final artistic rendering

The processing is optimized using NumPy's vectorized operations and OpenCV's efficient image processing capabilities.

Performance Optimizations

• Batch Processing: Videos with high FPS are processed in batches to improve throughput
• Resize Control: Large videos are automatically resized to Full HD for efficient processing
• Vectorized Operations: Uses NumPy's vectorized calculations instead of nested loops where possible
• Distance Matrix Calculation: Pre-computes distance matrices once per window size for speed
• Resize-based Average Calculation: Uses OpenCV's resize function to quickly compute average brightness values

Example Results

Image Processing

Original Image

Processed (window_size=5)

Processed (window_size=10)

Webcam Processing

Below is an example of the webcam processing effect.

License

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

Acknowledgments

• Inspired by traditional stippling art techniques
• Built with OpenCV and NumPy libraries