Vision 2.1.0

New Features 🥳

• Added a KMeans color clustering implementaion, and the following methods:
  • kmeansPosterize()
  • kmeansGroupImageColors()
• Vastly improved ArrayTools with many convenience methods (distinct, average, median...)
• Added image hashing capabilities, now supports ahash (average hash) and phash (perceptual hash)
• Check The Changelog For More Details

Bugfixes & Tweaks 🛠️

• Fixed Image.paintPixel regression where alpha channel wasn't considered correctly
• Added a ByteArray.from overload for Int64
• Fixed Radix not supporting different number types correctly
• Fixed Cramer not considering vision_quiet, and added descriptive exceptions when using the class
• Check The Changelog For More Details

Vision 2.0.0

New Features 🥳

• Added projective & affine transformation methods, as well as the following classes:
  • Matrix2D
  • TransformationMatrix2D
• Added different distortion methods: pincushionDistortion, barrelDistortion, fisheyeDistortion...
• Added 5 different color distortion methods: smooth, posterize, sepia, pixelate, vignette
• Check The Changelog For More Details

Bugfixes & Tweaks 🛠️

• Fixed typo throughout source code: Kernal -> Kernel
• Heavily extended the helper functions of ByteArray, Array2D.
• Added 50+ predefined colors to the Color class.
• MathTools was expanded and split into two classes: itself and ArrayTools
• Image class was improved, with some missing functions implemented and many bugs fixed. Many convinience functions were added as well.
• Check The Changelog For More Details

Vision 1.3.1

New Features 🥳

• Added distanceTo() method to Point2D & IntPoint2D
• Check The Changelog For More Details

Bugfixes & Tweaks 🛠️

• Fixed compile error on c++, cppia & c#
• Improved SimpleLineDetector's line filtering
• Check The Changelog For More Details

Vision 1.3.0

New Features 🥳

• Added Bitwise operators |, & and ^, you can now perform per-pixel operation on entire images at once!
• Added Image.fromBytes() and Image.toBytes()
• Added abstracts Int16Point2D & UInt16Point2D, useful for squeezing out a bit more performance 😉
• Added PixelFormat class - allows easy manipulation of byte arrays of different color formats
• Check The Changelog For More Details

Bugfixes & Tweaks 🛠️

• Tweaked Vision.simpleLine2DDetection(), should now report less overlapping lines
• Fixed off-by-one inaccuracies with image views
• Deprecated Gaussian.createXxXKernal() and Gaussian.create2DKernalOfSize() in favor of Gaussian.createKernalOfSize()
• Check The Changelog For More Details

(yes, I know I missed some updates before this one, but Github doesn't let you make releases for older commits, so :( )

Vision 1.2.1

New Features 🥳

• Added from/to casting methods for JS Canvas & Image elements and HaxeUI's Image element and ImageData data type
• Added MathTools.cropDecimal()
• Check The Changelog For More Details

Bugfixes & Tweaks 🛠️

• Slightly optimized Vision.simpleLine2DDetection
• Fixed LibraryRequired errors being inaccurate on Ceramic
• Check The Changelog For More Details

Vision 1.2.0

New Features 🥳:

• Image Views - Manipulate Portions Of The Image, Without Copy-Pasting!
• View Manipulation methods (Get, Set, Has & Shapes!)
• More Methods For Floating Pixel Manipulation
• Added Color.blackOrWhite()
• Check The Changelog For More Details

Bugfixes & Tweaks 🛠️

• Fixed image.resize() With NearestNeighbor
• Pixel Iteration Has Been Optimized
• Fixed Image.paintPixel() Resetting alpha To 1
• Changed Color.grayscale()'s Signature To Accept An Optional, Extra Parameter
• Changed Image's Byte Offset From 4 To 11
• Check The Changelog For More Details

Vision 1.1.1

Bugfixes & Tweaks 🛠️

• Fixed inability to compile when vision_allow_resize is defined
• Fixed forced format dependency
• Optimized MathTools.boundInt
• Fixed MathTools.clamp()
• Check The Changelog For More Details

Vision 1.1.0

New Features 🥳:

• Erosion & Dilation
• Laplacian Edge Detection & Operator
• Robert's Cross Operator
• Bilateral Denoising
• Noising Functions (Salt&Pepper, Drop-Out, White Noise)
• Image Combination
• Check The Changelog For More Details

Bugfixes & Tweaks 🛠️

• Added full ImageTools.loadFromFile support on sys targets
• Fixed MathTools.average, MathTools.median For C#
• Fixed VisionThread preventing compilation on some targets
• Sped Up Image.clone()
• Sped up perwittEdgeDiffOperator()
• Check The Changelog For More Details

Vision 1.0.0

Finally, after nearly 4 months of work, I'm releasing the first official version of Vision 🥳

Currently, Vision offers:

• Edge Detection (perwitt, sobel, canny)
• Line Detection (simple, hough transform partially available)
• Ridge Detection (convolution)
• Edge Highlighting (perwitt, sobel)
• Blurring (nearest-neighbor, gaussian, median)
• Sharpening (convolution, deepfry, contrast)
• Normalization (normalize, limitColorRanges, replaceColorRanges)
• General Manipulation (grayscale, black&white, invert)
• Cross platform, cross framework Image class. For more information, visit

And that's just the beginning...