Halftone Vision

Halftone Vision is a CUDA-accelerated tool that applies artistic halftone effects to images and videos. Using both black & white and color modes, this project leverages NVIDIA CUDA and OpenCV to generate halftone effects in real time as well as for offline processing.

Features

• CUDA Acceleration: Uses NVIDIA CUDA to dramatically speed up halftone processing for high-resolution images and videos.
• Dual Modes:
  • Black & White Mode: Generates a classic halftone effect by varying dot sizes based on brightness.
  • Color Mode: Applies halftone effects while preserving color information by averaging colors in each block.
• Real-Time Processing: View the halftone effect in real time with interactive adjustments.
• Interactive Controls:
  • Trackbars: Adjust parameters such as "Dot Size" and "Mode (0:BW, 1:Color)".
  • Video Navigation: Seek through video frames using a "Frame" trackbar.
  • Keyboard Shortcuts: Toggle pause/play (P), step frame-by-frame (A for backward, D for forward), and save output.
• Offline Video Processing: Process and save the entire video offline with your selected parameters (by pressing S in video mode). This reprocesses the full video—not just a recording of the display—to produce a high-quality output.
• Automatic Downscaling: High-resolution content (e.g., 4K) is automatically downscaled to a maximum resolution of 1920×1080 to ensure compatibility with your display.

Requirements

• CUDA Toolkit: NVIDIA CUDA Toolkit must be installed and configured.
• OpenCV: OpenCV (version 4 or later recommended) must be installed.
• C++ Compiler: A CUDA-compatible compiler such as nvcc is required.

Compilation

Compile Halftone Vision using NVIDIA’s nvcc compiler. For example:

nvcc -std=c++11 -O2 halftone_vision.cu -o halftone_vision `pkg-config --cflags --libs opencv4`

Ensure that your pkg-config is properly configured for your OpenCV installation.

Usage

Image Mode

To process an image, run:

./halftone_vision myimage.jpg

Interactive Controls:

• Dot Size Trackbar: Adjusts the size of the halftone dots.
• Mode Trackbar: Switch between Black & White (0) and Color (1) modes.

Keyboard Shortcuts:

• S: Save the current processed image as halftone_output.png.
• ESC: Exit the application.

Image Demo

Watch the transformation in action! Below are the demo images:

• Original Video: image_in.png
• Halftone Output Video: halftone_output.png

These images showcase the original input and the artistic halftone effect applied by Halftone Vision.

Video Mode

To process a video, run:

./halftone_vision myvideo.mp4

Interactive Controls:

• Frame Trackbar: Seek through the video.
• Dot Size & Mode Trackbars: Adjust the halftone parameters in real time.

Keyboard Shortcuts:

• P: Toggle pause/play.
• A: Step backward one frame (when paused).
• D: Step forward one frame (when paused).
• S: Process and save the full video offline using the current parameters. This reprocesses the entire video and saves it as halftone_output.avi.
• ESC: Exit the application.

Video Demo

Watch the transformation in action! Below are the demo videos:

• Original Video: video.mp4
• Halftone Output Video: output.mp4

These videos showcase the original input and the artistic halftone effect applied by Halftone Vision.

How It Works

Halftone Vision divides the input image or video frame into blocks (cells). For each cell:

• In Black & White Mode, it computes the average brightness and draws a dot whose radius is proportional to the darkness of the cell.
• In Color Mode, it computes the average color and draws a colored dot accordingly.

CUDA kernels accelerate these computations, ensuring high-performance processing even for high-resolution content.

License

This project is open source and available under the MIT License.

Acknowledgments

• NVIDIA CUDA: For providing a powerful platform for GPU computing.
• OpenCV: For the robust computer vision framework used for image and video handling.

Happy halftoning!