OpenEXRimages_HighDynamicRange_Retina_toneMapping_video.cpp

//============================================================================
// Name        : OpenEXRimages_HighDynamicRange_Retina_toneMapping_video.cpp
// Author      : Alexandre Benoit (benoit.alexandre.vision@gmail.com)
// Version     : 0.2
// Copyright   : Alexandre Benoit, LISTIC Lab, december 2011
// Description : HighDynamicRange compression (tone mapping) for image sequences with the help of the Gipsa/Listic's retina in C++, Ansi-style
// Known issues: the input OpenEXR sequences can have bad computed pixels that should be removed
//               => a simple method consists of cutting histogram edges (a slider for this on the UI is provided)
//               => however, in image sequences, this histogramm cut must be done in an elegant way from frame to frame... still not done...     
//============================================================================

#include <iostream>
#include <stdio.h>
#include <cstring>

#include "opencv2/opencv.hpp"

void help(std::string errorMessage)
{
	std::cout<<"Program init error : "<<errorMessage<<std::endl;
	std::cout<<"\nProgram call procedure : ./OpenEXRimages_HighDynamicRange_Retina_toneMapping [OpenEXR image sequence to process] [OPTIONNAL start frame] [OPTIONNAL end frame]"<<std::endl;
	std::cout<<"\t[OpenEXR image sequence to process] : std::sprintf style ready prototype filename of the input HDR images to process, must be an OpenEXR format, see http://www.openexr.com/ to get some samples or create your own using camera bracketing and Photoshop or equivalent software for OpenEXR image synthesis"<<std::endl;
	std::cout<<"\t\t => WARNING : image index number of digits cannot exceed 10"<<std::endl;
	std::cout<<"\t[start frame] : the starting frame tat should be considered"<<std::endl;
	std::cout<<"\t[end frame] : the ending frame tat should be considered"<<std::endl;
	std::cout<<"\nExamples:"<<std::endl;
	std::cout<<"\t-Image processing : ./OpenEXRimages_HighDynamicRange_Retina_toneMapping_video memorial%3d.exr 20 45"<<std::endl;
	std::cout<<"\t-Image processing : ./OpenEXRimages_HighDynamicRange_Retina_toneMapping_video memorial%3d.exr 20 45 log"<<std::endl;
	std::cout<<"\t ==> to process images from memorial020d.exr to memorial045d.exr"<<std::endl;

}

// simple procedure for 1D curve tracing
void drawPlot(const cv::Mat curve, const std::string figureTitle, const int lowerLimit, const int upperLimit)
{
	//std::cout<<"curve size(h,w) = "<<curve.size().height<<", "<<curve.size().width<<std::endl;
	cv::Mat displayedCurveImage = cv::Mat::ones(200, curve.size().height, CV_8U);

	cv::Mat windowNormalizedCurve;
	normalize(curve, windowNormalizedCurve, 0, 200, CV_MINMAX, CV_32F);

	displayedCurveImage = cv::Scalar::all(255); // set a white background
	int binW = cvRound((double)displayedCurveImage.cols/curve.size().height);

	for( int i = 0; i < curve.size().height; i++ )
		rectangle( displayedCurveImage, cv::Point(i*binW, displayedCurveImage.rows),
				cv::Point((i+1)*binW, displayedCurveImage.rows - cvRound(windowNormalizedCurve.at<float>(i))),
				cv::Scalar::all(0), -1, 8, 0 );
	rectangle( displayedCurveImage, cv::Point(0, 0),
			cv::Point((lowerLimit)*binW, 200),
			cv::Scalar::all(128), -1, 8, 0 );
	rectangle( displayedCurveImage, cv::Point(displayedCurveImage.cols, 0),
			cv::Point((upperLimit)*binW, 200),
			cv::Scalar::all(128), -1, 8, 0 );

	cv::imshow(figureTitle, displayedCurveImage);
}

/*
 * objective : get the gray level map of the input image and rescale it to the range [0-255] if rescale0_255=TRUE, simply trunks else
 */
void rescaleGrayLevelMat(const cv::Mat &inputMat, cv::Mat &outputMat, const float histogramClippingLimit, const bool rescale0_255)
 {
	 // adjust output matrix wrt the input size but single channel
	 std::cout<<"Input image rescaling with histogram edges cutting (in order to eliminate bad pixels created during the HDR image creation) :"<<std::endl;
	 //std::cout<<"=> image size (h,w,channels) = "<<inputMat.size().height<<", "<<inputMat.size().width<<", "<<inputMat.channels()<<std::endl;
	 //std::cout<<"=> pixel coding (nbchannel, bytes per channel) = "<<inputMat.elemSize()/inputMat.elemSize1()<<", "<<inputMat.elemSize1()<<std::endl;

	 // get min and max values to use afterwards if no 0-255 rescaling is used
	 double maxInput, minInput, histNormRescalefactor=1.f;
	 double histNormOffset=0.f;
	 minMaxLoc(inputMat, &minInput, &maxInput);
	 histNormRescalefactor=255.f/(maxInput-minInput);
	 histNormOffset=minInput;
	 std::cout<<"Hist max,min = "<<maxInput<<", "<<minInput<<" => scale, offset = "<<histNormRescalefactor<<", "<<histNormOffset<<std::endl;
	 // rescale between 0-255, keeping floating point values
	 cv::Mat normalisedImage;
	 cv::normalize(inputMat, normalisedImage, 0.f, 255.f, cv::NORM_MINMAX);
	 if (rescale0_255)
		normalisedImage.copyTo(outputMat);
	 // extract a 8bit image that will be used for histogram edge cut
	 cv::Mat intGrayImage;
	 if (inputMat.channels()==1)
	 {
		 normalisedImage.convertTo(intGrayImage, CV_8U);
	 }else
	 {
		 cv::Mat rgbIntImg;
		 normalisedImage.convertTo(rgbIntImg, CV_8UC3);
		 cvtColor(rgbIntImg, intGrayImage, CV_BGR2GRAY);
	 }

	 // get histogram density probability in order to cut values under above edges limits (here 5-95%)... usefull for HDR pixel errors cancellation
	 cv::Mat dst, hist;
	 int histSize = 256;
	 calcHist(&intGrayImage, 1, 0, cv::Mat(), hist, 1, &histSize, 0);
	 cv::Mat normalizedHist;
	
	 normalize(hist, normalizedHist, 1.f, 0.f, cv::NORM_L1, CV_32F); // normalize histogram so that its sum equals 1

	 // compute density probability
	 cv::Mat denseProb=cv::Mat::zeros(normalizedHist.size(), CV_32F);
	 denseProb.at<float>(0)=normalizedHist.at<float>(0);
	 int histLowerLimit=0, histUpperLimit=0;
	 for (int i=1;i<normalizedHist.size().height;++i)
	 {
		 denseProb.at<float>(i)=denseProb.at<float>(i-1)+normalizedHist.at<float>(i);
		 //std::cout<<normalizedHist.at<float>(i)<<", "<<denseProb.at<float>(i)<<std::endl;
		 if ( denseProb.at<float>(i)<histogramClippingLimit)
			 histLowerLimit=i;
		 if ( denseProb.at<float>(i)<1.f-histogramClippingLimit)
			 histUpperLimit=i;
	 }
	 // deduce min and max admitted gray levels
	 float minInputValue = (float)histLowerLimit/histSize*255.f;
	 float maxInputValue = (float)histUpperLimit/histSize*255.f;

	 std::cout<<"=> Histogram limits "
			 <<"\n\t"<<histogramClippingLimit*100.f<<"% index = "<<histLowerLimit<<" => normalizedHist value = "<<denseProb.at<float>(histLowerLimit)<<" => input gray level = "<<minInputValue
			 <<"\n\t"<<(1.f-histogramClippingLimit)*100.f<<"% index = "<<histUpperLimit<<" => normalizedHist value = "<<denseProb.at<float>(histUpperLimit)<<" => input gray level = "<<maxInputValue
			 <<std::endl;
	 //drawPlot(denseProb, "input histogram density probability", histLowerLimit, histUpperLimit);
	 drawPlot(normalizedHist, "input histogram", histLowerLimit, histUpperLimit);

	if(rescale0_255) // rescale between 0-255 if asked to
	{
		cv::threshold( outputMat, outputMat, maxInputValue, maxInputValue, 2 ); //THRESH_TRUNC, clips values above maxInputValue
		cv::threshold( outputMat, outputMat, minInputValue, minInputValue, 3 ); //THRESH_TOZERO, clips values under minInputValue
		// rescale image range [minInputValue-maxInputValue] to [0-255]
		outputMat-=minInputValue;
		outputMat*=255.f/(maxInputValue-minInputValue);
	}else
	{
		inputMat.copyTo(outputMat);
		// update threshold in the initial input image range
		maxInputValue=(maxInputValue-255.f)/histNormRescalefactor+maxInput;
		minInputValue=minInputValue/histNormRescalefactor+minInput;
	 	std::cout<<"===> Input Hist clipping values (max,min) = "<<maxInputValue<<", "<<minInputValue<<std::endl;
		cv::threshold( outputMat, outputMat, maxInputValue, maxInputValue, 2 ); //THRESH_TRUNC, clips values above maxInputValue
		cv::threshold( outputMat, outputMat, minInputValue, minInputValue, 3 ); //
	}
 }

 // basic callback method for interface management
 cv::Mat inputImage;
 cv::Mat imageInputRescaled;
 float globalRescalefactor=1;
 cv::Scalar globalOffset=0;
 int histogramClippingValue;
 void callBack_rescaleGrayLevelMat(int, void*)
 {
	 std::cout<<"Histogram clipping value changed, current value = "<<histogramClippingValue<<std::endl;
	// rescale and process
	inputImage+=globalOffset;
	inputImage*=globalRescalefactor;
	inputImage+=cv::Scalar(50, 50, 50, 50); // WARNING value linked to the hardcoded value (200.0) used in the globalRescalefactor in order to center on the 128 mean value... experimental but... basic compromise
	rescaleGrayLevelMat(inputImage, imageInputRescaled, (float)histogramClippingValue/100.f, true);

 }

 cv::Ptr<cv::Retina> retina;
 int retinaHcellsGain;
 int localAdaptation_photoreceptors, localAdaptation_Gcells;
 void callBack_updateRetinaParams(int, void*)
 {

	 retina->setupOPLandIPLParvoChannel(true, true, (float)(localAdaptation_photoreceptors/200.0), 0.5f, 0.43f, (double)retinaHcellsGain, 1.f, 7.f, (float)(localAdaptation_Gcells/200.0));
 }

 int colorSaturationFactor;
 void callback_saturateColors(int, void*)
 {
	 retina->setColorSaturation(true, (float)colorSaturationFactor);
 }

// loadNewFrame : loads a n image wrt filename parameters. it also manages image rescaling/histogram edges cutting (acts differently at first image i.e. if firstTimeread=true)
void loadNewFrame(const std::string filenamePrototype, const int currentFileIndex, const bool firstTimeread)
{
	 char *currentImageName=NULL;
	currentImageName = (char*)malloc(sizeof(char)*filenamePrototype.size()+10);

	// grab the first frame	 
	sprintf(currentImageName, filenamePrototype.c_str(), currentFileIndex);
	
	 //////////////////////////////////////////////////////////////////////////////
	 // checking input media type (still image, video file, live video acquisition)
	 std::cout<<"RetinaDemo: reading image : "<<currentImageName<<std::endl;
	 // image processing case
	 // declare the retina input buffer... that will be fed differently in regard of the input media
	 inputImage = cv::imread(currentImageName, -1); // load image in RGB mode
	 std::cout<<"=> image size (h,w) = "<<inputImage.size().height<<", "<<inputImage.size().width<<std::endl;
	 if (inputImage.empty())
	 {
	    help("could not load image, program end");
            return;; 
         }

	// rescaling/histogram clipping stage
	// rescale between 0 and 1
	// TODO : take care of this step !!! maybe disable of do this in a nicer way ... each successive image should get the same transformation... but it depends on the initial image format
	double maxInput, minInput;
	minMaxLoc(inputImage, &minInput, &maxInput);
	std::cout<<"ORIGINAL IMAGE pixels values range (max,min) : "<<maxInput<<", "<<minInput<<std::endl
;if (firstTimeread)
	{
		/* the first time, get the pixel values range and rougthly update scaling value
		in order to center values around 128 and getting a range close to [0-255], 
		=> actually using a little less in order to let some more flexibility in range evolves...
		*/
		double maxInput, minInput;
	 	minMaxLoc(inputImage, &minInput, &maxInput);
	 	std::cout<<"FIRST IMAGE pixels values range (max,min) : "<<maxInput<<", "<<minInput<<std::endl;
		globalRescalefactor=50.0/(maxInput-minInput); // less than 255 for flexibility... experimental value to be carefull about
		float channelOffset = -1.5*minInput;
		globalOffset= cv::Scalar(channelOffset, channelOffset, channelOffset, channelOffset);	
	}
	// call the generic input image rescaling callback	
	callBack_rescaleGrayLevelMat(1,NULL);
}

 int main(int argc, char* argv[]) {
	 // welcome message
	 std::cout<<"*********************************************************************************"<<std::endl;
	 std::cout<<"* Retina demonstration for High Dynamic Range compression (tone-mapping) : demonstrates the use of a wrapper class of the Gipsa/Listic Labs retina model."<<std::endl;
	 std::cout<<"* This retina model allows spatio-temporal image processing (applied on still images, video sequences)."<<std::endl;
	 std::cout<<"* This demo focuses demonstration of the dynamic compression capabilities of the model"<<std::endl;
	 std::cout<<"* => the main application is tone mapping of HDR images (i.e. see on a 8bit display a more than 8bits coded (up to 16bits) image with details in high and low luminance ranges"<<std::endl;
	 std::cout<<"* The retina model still have the following properties:"<<std::endl;
	 std::cout<<"* => It applies a spectral whithening (mid-frequency details enhancement)"<<std::endl;
	 std::cout<<"* => high frequency spatio-temporal noise reduction"<<std::endl;
	 std::cout<<"* => low frequency luminance to be reduced (luminance range compression)"<<std::endl;
	 std::cout<<"* => local logarithmic luminance compression allows details to be enhanced in low light conditions\n"<<std::endl;
	 std::cout<<"* for more information, reer to the following papers :"<<std::endl;
	 std::cout<<"* Benoit A., Caplier A., Durette B., Herault, J., \"USING HUMAN VISUAL SYSTEM MODELING FOR BIO-INSPIRED LOW LEVEL IMAGE PROCESSING\", Elsevier, Computer Vision and Image Understanding 114 (2010), pp. 758-773, DOI: http://dx.doi.org/10.1016/j.cviu.2010.01.011"<<std::endl;
	 std::cout<<"* Vision: Images, Signals and Neural Networks: Models of Neural Processing in Visual Perception (Progress in Neural Processing),By: Jeanny Herault, ISBN: 9814273686. WAPI (Tower ID): 113266891."<<std::endl;
	 std::cout<<"* => reports comments/remarks at benoit.alexandre.vision@gmail.com"<<std::endl;
	 std::cout<<"* => more informations and papers at : http://sites.google.com/site/benoitalexandrevision/"<<std::endl;
	 std::cout<<"*********************************************************************************"<<std::endl;
	 std::cout<<"** WARNING : this sample requires OpenCV to be configured with OpenEXR support **"<<std::endl;
	 std::cout<<"*********************************************************************************"<<std::endl;
	 std::cout<<"*** You can use free tools to generate OpenEXR images from images sets   :    ***"<<std::endl;
	 std::cout<<"*** =>  1. take a set of photos from the same viewpoint using bracketing      ***"<<std::endl;
	 std::cout<<"*** =>  2. generate an OpenEXR image with tools like qtpfsgui.sourceforge.net ***"<<std::endl;
	 std::cout<<"*** =>  3. apply tone mapping with this program                               ***"<<std::endl;
	 std::cout<<"*********************************************************************************"<<std::endl;

	 // basic input arguments checking
	 if (argc<4)
	 {
		 help("bad number of parameter");
		 return -1;
	 }

	 bool useLogSampling = !strcmp(argv[argc-1], "log"); // check if user wants retina log sampling processing

	 int startFrameIndex=0, endFrameIndex=0, currentFrameIndex=0;
	 sscanf(argv[2], "%d", &startFrameIndex);
	 sscanf(argv[3], "%d", &endFrameIndex);
	 std::string inputImageNamePrototype(argv[1]);

	 //////////////////////////////////////////////////////////////////////////////
	 // checking input media type (still image, video file, live video acquisition)
	 std::cout<<"RetinaDemo: setting up system with first image..."<<std::endl;
	 loadNewFrame(inputImageNamePrototype, startFrameIndex, true);

	 if (inputImage.empty())
	 {
	    help("could not load image, program end");
            return -1; 
         }

	 //////////////////////////////////////////////////////////////////////////////
	 // Program start in a try/catch safety context (Retina may throw errors)
	 try
	 {
		 /* create a retina instance with default parameters setup, uncomment the initialisation you wanna test
		  * -> if the last parameter is 'log', then activate log sampling (favour foveal vision and subsamples peripheral vision)
		  */
		 if (useLogSampling)
                {
                     retina = new cv::Retina(inputImage.size(),true, cv::RETINA_COLOR_BAYER, true, 2.0, 10.0);
                 }
		 else// -> else allocate "classical" retina :
		     retina = new cv::Retina(inputImage.size());
		
		// save default retina parameters file in order to let you see this and maybe modify it and reload using method "setup"
		retina->write("RetinaDefaultParameters.xml");

                 // desactivate Magnocellular pathway processing (motion information extraction) since it is not usefull here
                 retina->activateMovingContoursProcessing(false);

		 // declare retina output buffers
		 cv::Mat retinaOutput_parvo;

		 /////////////////////////////////////////////
		 // prepare displays and interactions
		 histogramClippingValue=0; // default value... updated with interface slider

		 std::string retinaInputCorrected("Retina input image (with cut edges histogram for basic pixels error avoidance)");
		 cv::namedWindow(retinaInputCorrected,1);
		 cv::createTrackbar("histogram edges clipping limit", "Retina input image (with cut edges histogram for basic pixels error avoidance)",&histogramClippingValue,50,callBack_rescaleGrayLevelMat);

		 std::string RetinaParvoWindow("Retina Parvocellular pathway output : 16bit=>8bit image retina tonemapping");
		 cv::namedWindow(RetinaParvoWindow, 1);
		 colorSaturationFactor=3;
		 cv::createTrackbar("Color saturation", "Retina Parvocellular pathway output : 16bit=>8bit image retina tonemapping", &colorSaturationFactor,5,callback_saturateColors);

		 retinaHcellsGain=40;
		 cv::createTrackbar("Hcells gain", "Retina Parvocellular pathway output : 16bit=>8bit image retina tonemapping",&retinaHcellsGain,100,callBack_updateRetinaParams);

		 localAdaptation_photoreceptors=197;
		 localAdaptation_Gcells=190;
		 cv::createTrackbar("Ph sensitivity", "Retina Parvocellular pathway output : 16bit=>8bit image retina tonemapping", &localAdaptation_photoreceptors,199,callBack_updateRetinaParams);
		 cv::createTrackbar("Gcells sensitivity", "Retina Parvocellular pathway output : 16bit=>8bit image retina tonemapping", &localAdaptation_Gcells,199,callBack_updateRetinaParams);

		std::string powerTransformedInput("EXR image with basic processing : 16bits=>8bits with gamma correction");

		 /////////////////////////////////////////////
		 // apply default parameters of user interaction variables
		 callBack_updateRetinaParams(1,NULL); // first call for default parameters setup
		 callback_saturateColors(1, NULL);

		 // processing loop with stop condition
		 currentFrameIndex=startFrameIndex;
		 while(currentFrameIndex <= endFrameIndex)
		 {
			 loadNewFrame(inputImageNamePrototype, currentFrameIndex, false);

			 if (inputImage.empty())
			 {
			    std::cout<<"Could not load new image (index = "<<currentFrameIndex<<"), program end"<<std::endl;
			    return -1; 
			 }
			// display input & process standard power transformation 	
 			imshow("EXR image original image, 16bits=>8bits linear rescaling ", imageInputRescaled);
	 		cv::Mat gammaTransformedImage;
	 		cv::pow(imageInputRescaled, 1./5, gammaTransformedImage); // apply gamma curve: img = img ** (1./5)
	 		imshow(powerTransformedInput, gammaTransformedImage);
			 // run retina filter
			 retina->run(imageInputRescaled);
			 // Retrieve and display retina output
			 retina->getParvo(retinaOutput_parvo);
			 cv::imshow(retinaInputCorrected, imageInputRescaled/255.f);
			 cv::imshow(RetinaParvoWindow, retinaOutput_parvo);
			 cv::waitKey(4);
			// jump to next frame			
			++currentFrameIndex;
		 }
	 }catch(cv::Exception e)
	 {
		 std::cerr<<"Error using Retina : "<<e.what()<<std::endl;
	 }

	 // Program end message
	 std::cout<<"Retina demo end"<<std::endl;

	 return 0;
 }