Fast and Resolution Independent Line Integral Convolution (FastLIC) in Java

This project is an implementation of the Fast and Resolution Independent Line Integral Convolution (LIC) algorithm for flow visualization, written entirely in Java. The LIC techniques are widely used for visualizing vector fields, generating high-quality textures that represent the flow direction in the field. This project also includes the original LIC implementation.

Introduction

LIC is a method for visualizing vector fields by convolving a noise texture along streamlines of the field, creating streak-like patterns that intuitively represent the direction and structure of the flow [1]. FastLIC is a variation designed to speed up image generation by applying Gaussian blur filtering [2].

The basic workflow of this project includes:

• Input Texture: You can provide an input image (JPG, grayscale recommended). If not provided, the system will automatically generate a white noise texture.
• Vector Field Synthesis: The implementation supports multiple vector field types:
  • SADDLE (hyperbolic flow)
  • CENTER (circular flow)
  • SOURCE (radial outward flow)
  • SINK (radial inward flow)
  • SPIRAL_SOURCE (spiral outward flow)
  • SPIRAL_SINK (spiral inward flow)
• Fast Line Integral Convolution: Applies the FastLIC algorithm to the chosen vector field and input texture, generating a smooth streaked texture that visually represents the flow.
• Output Image: The resulting image is saved as a JPG file.

Usage Instructions

Requirements

• Java 8 or newer;
• No external dependencies (pure Java, uses standard libraries).

How to Run

1. Clone this repository.

2. Place any input image (JPG format) in the src/main/resources folder if you want to use a custom texture.

3. Example usage in Java:

public class Main {
    public static void main(String[] args) throws IOException {
        FastLICBuilder builder = new FastLICBuilder()
                .setFlowFieldType(FlowField.Type.SADDLE)
                .setSquareFlowFieldSize(400)
                .setDiscreteFilterSize(2048)
                .setLowPassFilterLength(10.0f)
                .setLineSquareClipMax(100000.0f)
                .setVectorComponentMinimum(0.05f);

        // Optionally, set a custom input image from resources
        // builder.setInputImage("noise.jpg");

        builder.generate("fastlic_output.jpg");
    }
}

4. Output images are saved in the folder: target/output/.

Supported Vector Fields

You can select the vector field type with:

builder.setFlowFieldType(FlowField.Type.TYPE);

Available types:

Type	Description
SADDLE	Hyperbolic flow (saddle point)
CENTER	Circular flow around the center
SOURCE	Radial outward flow
SINK	Radial inward flow
SPIRAL_SOURCE	Spiral outward flow
SPIRAL_SINK	Spiral inward flow

Project Structure

• FastLICBuilder.java — Main class to configure and run the FastLIC algorithm.
• FlowField.java — Defines vector field types and generation methods.
• Vector2D.java — Simple 2D vector class with normalization.
• Main.java — Example entry point.

Notes

• Only JPG format is supported for input and output images.
• Input images are expected to be in grayscale; if not, the program converts them by extracting the red channel.

License

The available source codes here are under the Apache License, version 2.0 (see the attached LICENSE file for more details). Any questions can be submitted to my email: [email protected].

References

[1] B. Cabral and L. Leedom, "Imaging Vector Fields Using Line Integral Convolution", Proceedings of SIGGRAPH '93. [2] D. Stalling and H. C. Hege, "Fast and Resolution Independent Line Integral Convolution", Proceedings of ACM SIGGRAPH '95.