FastImage — SIMD‑Accelerated, Off‑Heap Image Processing for Java

Ultra-fast native image processing using AVX2/SSE4.1 kernels and zero-GC memory management.

FastImage ist eine ultra‑schnelle, native‑beschleunigte Image‑Processing‑Engine für Java, gebaut für das FastJava‑Ecosystem. Es kombiniert AVX/SSE SIMD, off‑heap Storage, zero‑copy Pipelines und eine fluent API, um typische BufferedImage‑Operationen 10–50× schneller auszuführen — ohne GC‑Pressure, ohne Pixel‑Loops, ohne JVM‑Overhead.

// Quick Start — SIMD-Accelerated Filtering
import fastimage.FastImage;

public class Demo {
    public static void main(String[] args) {
        FastImage img = FastImage.load("input.jpg");
        
        img.adjustContrast(1.2f)
           .blurStack(15.0f)
           .grayscale();
           
        img.save("output.png");
        img.dispose(); // Free native memory
    }
}

---

Table of Contents

• Key Features
• Performance
• Installation
• Try the Demo
• API Reference
• Platform Support
• Building from Source
• License
• Related Projects

---

Key Features

• 🚀 SIMD Acceleration — Hand-optimized C++ kernels using AVX2 and SSE4.1 vector instructions.
• 🧠 Zero-GC Overhead — Pixels are stored in off-heap memory, preventing GC pauses during heavy manipulation.
• 🌀 Advanced Blur Suite — Real-time Gaussian, Stack (iOS-style), and Kawase blurs.
• 🛡️ Fail-Safe JNI — Robust error handling with FastImageException and native handle validation.
• 🔄 Fast Conversion — Optimized bit-copying between BufferedImage and native memory.

---

Performance

FastImage utilizes the full power of your CPU, outperforming standard Java2D loops by orders of magnitude:

Operation	Java2D (BufferedImage)	FastImage (SIMD)	Speedup
Brightness	~48.6 ms/op	~1.5 ms/op	32x
Gaussian Blur (r10)	~1100.0 ms/op	~170.4 ms/op	6.5x
Grayscale	~20.0 ms/op	~1.3 ms/op	15x

Tested on: 1920x1080 (1080p) ARGB Image on Intel i7-12700K.

---

Installation

FastJava modules require two dependencies: the module itself, and FastCore (which handles the native library extraction).

Maven (JitPack)

<repositories>
    <repository>
        <id>jitpack.io</id>
        <url>https://jitpack.io</url>
    </repository>
</repositories>

<dependencies>
    <dependency>
        <groupId>com.github.andrestubbe</groupId>
        <artifactId>fastimage</artifactId>
        <version>0.1.0</version>
    </dependency>
    <dependency>
        <groupId>com.github.andrestubbe</groupId>
        <artifactId>fastcore</artifactId>
        <version>0.1.0</version>
    </dependency>
</dependencies>

Gradle (JitPack)

repositories {
    maven { url 'https://jitpack.io' }
}

dependencies {
    implementation 'com.github.andrestubbe:fastimage:0.1.0'
    implementation 'com.github.andrestubbe:fastcore:0.1.0'
}

Option 3: Direct Download

Download the latest JARs directly from the releases:

1. 📦 fastimage-v0.1.0.jar
2. ⚙️ fastcore-v0.1.0.jar

---

Try the Demo

1. Clone this repository: git clone https://github.com/andrestubbe/FastImage.git
2. Run the automated showcase: .\run-demo.bat

Includes the interactive Visual Editor and the Blur Gallery.

---

API Reference

Method	Description
void grayscale()	Converts image to luminance-weighted grayscale via SIMD.
void adjustBrightness(f)	Scales RGB values with saturation clamping.
void adjustContrast(f)	Adjusts image contrast around the midpoint.
void blurGaussian(r)	High-quality Gaussian blur approximation ($O(N)$).
void blurStack(r)	Extremely fast separable weighted blur.
void resize(w, h)	Bilinear resizing using native kernels.

---

Platform Support

Architecture	Instruction Set	OS
x64	AVX2 (Runtime Dispatch)	Windows 10/11
x64	SSE4.1 (Fallback)	Windows 10/11

---

Building from Source

For detailed instructions on compiling the C++ JNI code and building the Maven FatJAR, see COMPILE.md.

---

License

MIT License — See LICENSE file for details.

---

Related Projects

• FastCore — Native Library Loader
• FastTheme — Native Window Styling
• FastGraphics — Hardware-accelerated 2D Rendering

---

Made with ⚡ by Andre Stubbe