kanryu/luminor

A simple and fast library that changes the brightness, contrast, and gamma value of on-memory images in real time generated with Halide

Luminor

A simple and fast library that changes the brightness, contrast, and gamma value of on-memory images in real time generated with Halide language.

Summary

• Simple C/C++ language API
• Optimized processing automatically generated with Halide language
  • The processing is completed in approximately 100 of the time of the normal implementation.
    • Measured with [email protected] GHz (built by gcc-7.2.0 CFLAGS=-O3)
    • 100 times without LUT, 2.9 times with LUT
    • e.g. https://github.com/FFmpeg/FFmpeg/blob/master/libavfilter/vf_eq.c
• It consists of C/C++ header file and static library

This is a part of QuickViewer.

Binary Releases

https://github.com/kanryu/luminor/releases

API

int luminor(struct halide_buffer_t *_input_buffer, float _brightness, float _contrast, float _gamma, struct halide_buffer_t *_output_buffer)

• _input_buffer
  • input bitmap of rgb24 or gray8
  • gray8 should be initialized as 3D Buffer(width, height, 1)
• _brightness
  • Add the value to each pixel
  • default: 0.0, min: -128.0, max: 127.0
• _contrast
  • Multiply each pixel by the value(the base value is 0.5(128))
  • default: 1.0, min: 0.01, max: 10.0
• _gamma
  • Change the gamma value of each pixel
  • It gets brighter when it gets bigger, it gets darker when it gets smaller
  • default: 1.0, min: 0.01, max: 10.0
• _output_buffer
  • output bitmap, the same pixel format as _input_buffer

int luminor_rgba(struct halide_buffer_t *_input_buffer, float _brightness, float _contrast, float _gamma, struct halide_buffer_t *_output_buffer)

• _input_buffer
  • input bitmap of rgba32
  • alpha byte must be at 4th byte of each pixel
  • need to call Buffer::make_interleaved() to prepare buffers

How to use

Only to include luminor.h, HalideBuffer.h and link luminor.a, the API will work.

Details to see https://github.com/kanryu/quickviewer/blob/master/luminor/qluminor.cpp

How to build

You have to get Halide. If you just build the library binary version of Halide is enough. However, if you want to check the operation of the sample program, also download Halide source code.

And,

$ make

Files

• images/
  • test images
• msvc/
  • Makefile for nmake
• msvc_sln/
  • solution/project for Visual Studio 2015
• py/
  • for Python3 and python_bindings for Halide language
• luminor_gen.cpp
  • library generator for luminor(), luminor_rgba()
• samplemain.cpp
  • sample program for luminor/luninor_rgba
  • currently luminor_rgba is not supported

License

MIT

Author

Copyright (c) 2017, KATO Kanryu([email protected])