pockethandbook.h

#pragma once
#ifndef POCKETHANDBOOK_H__INCLUDED
#define POCKETHANDBOOK_H__INCLUDED

#include <Halide.h>
#include <string>

namespace PocketHandbook {

// set of test suites for PocketHandbook algorithms
class TestSuite {
public:
	static void Run() {
		TestHSI();
		RunHSIExample();
		/*
		TestGeometricMean();
		RunGeometricMeanExample();
		TestFlip(); 
		RunFlipExample();
		TestBinaryDilationFilter();
		RunBinaryDilationOnRGBImage();
		TestBrightness(); 
		TestGraylevelHistogram();
		*/
	}
private:
	static void TestGraylevelHistogram();
	static void TestBrightness();
	static void TestBinaryDilationFilter();
	static void RunBinaryDilationOnRGBImage();
	static void TestFlip();
	static void RunFlipExample();
	static void TestGeometricMean();
	static void RunGeometricMeanExample();
	static void TestHSI();
	static void RunHSIExample();
};
	
// compute the "gray level histogram" of an image.. NOTE:
// in all cases, we'll assume RGB images instead of grayscale and just
// compute the "gray level" for each channel individually
class GraylevelHistogram
{
	bool _validate;
	
public:
	GraylevelHistogram()
		: _validate(false)
	{
	}
	
	// compute the histogram of the input 3 channel image and return a 256x3 output
	// results as the normalized histogram. 
	// Output Dimensions:
	// [0] = gray level
	// [1] = color channel
	// Values = sum of pixels at this gray level [0..255] divided by total number of pixels in the image
	Halide::Image<float> Run(Halide::Image<uint8_t> input);
	
	// convert the normalized histogram from Run into an image
	void SaveHistogramToImage(Halide::Image<float> histogram, std::string filename);
};

// Increase or decrease the brightness of an image
class Brightness
{
public:
	Halide::Image<uint8_t> Run(Halide::Image<uint8_t> input, int brightness);
};

// Dilate the image with a structuring function
class BinaryDilationFilter
{
public:
	Halide::Image<uint8_t> Run(Halide::Image<uint8_t> input);
};

// Flip an image
// pg. 80
class Flip
{
public:
	// if upDown == false, flips the image left-right
	Halide::Image<uint8_t> Run(Halide::Image<uint8_t> input, bool upDown);
};

// Compute the geometric mean of an image
// pg. 88
class GeometricMeanFilter
{
public:
	// N = size of region, must be an odd number less than 12
	Halide::Image<uint8_t> Run(Halide::Image<uint8_t> input, int N);
};

// HSI - Hue Saturation Intensity Color Model
// pg. 97-98
class HSI
{
public:
	// The resulting 3 channels are H, S, and I
	Halide::Image<uint8_t> Run(Halide::Image<uint8_t> input);
};

}

#endif //POCKETHANDBOOK_H__INCLUDED