Example 1: Tsukuba
 Left Image |
 Right Image |
 SAD Disparity Map
Disparity Range: 0-16
Window Size: 9×9
The hotter the color, the closer it is; the cooler the color the farther it is. |
Example 2: Stereogram
 Left Image |
 Right Image |
 SAD Disparity Map
Disparity Range: 0-16
Window Size: 9×9
The hotter the color, the closer it is; the cooler the color the farther it is. |
MATLAB Code-SAD Left to Right Matching-Implementation 1
% *************************************************************************
% Title: Function-Compute Correlation between two images using the
% similarity measure of Sum of Absolute Differences (SAD) with Left Image
% as reference.
% Author: Siddhant Ahuja
% Created: May 2008
% Copyright Siddhant Ahuja, 2008
% Inputs: Left Image (var: leftImage), Right Image (var: rightImage),
% Window Size (var: windowSize), Minimum Disparity (dispMin), Maximum
% Disparity (dispMax)
% Outputs: Disparity Map (var: dispMap), Time taken (var: timeTaken)
% Example Usage of Function: [dispMap, timeTaken]=funcSADL2R('StereogramLeft.jpg', 'StereogramRight.jpg', 9, 0, 16);
% *************************************************************************
function [dispMap, timeTaken]=funcSADL2R(leftImage, rightImage, windowSize, dispMin, dispMax)
try
% Grab the image information (metadata) of left image using the function imfinfo
leftImageInfo=imfinfo(leftImage);
% Since SADL2R is applied on a grayscale image, determine if the
% input left image is already in grayscale or color
if(getfield(leftImageInfo,'ColorType')=='truecolor')
% Read an image using imread function, convert from RGB color space to
% grayscale using rgb2gray function and assign it to variable leftImage
leftImage=rgb2gray(imread(leftImage));
% Convert the image from uint8 to double
leftImage=double(leftImage);
else if(getfield(leftImageInfo,'ColorType')=='grayscale')
% If the image is already in grayscale, then just read it.
leftImage=imread(leftImage);
% Convert the image from uint8 to double
leftImage=double(leftImage);
else
error('The Color Type of Left Image is not acceptable. Acceptable color types are truecolor or grayscale.');
end
end
catch
% if it is not an image but a variable
leftImage=leftImage;
end
try
% Grab the image information (metadata) of right image using the function imfinfo
rightImageInfo=imfinfo(rightImage);
% Since SADL2R is applied on a grayscale image, determine if the
% input right image is already in grayscale or color
if(getfield(rightImageInfo,'ColorType')=='truecolor')
% Read an image using imread function, convert from RGB color space to
% grayscale using rgb2gray function and assign it to variable rightImage
rightImage=rgb2gray(imread(rightImage));
% Convert the image from uint8 to double
rightImage=double(rightImage);
else if(getfield(rightImageInfo,'ColorType')=='grayscale')
% If the image is already in grayscale, then just read it.
rightImage=imread(rightImage);
% Convert the image from uint8 to double
rightImage=double(rightImage);
else
error('The Color Type of Right Image is not acceptable. Acceptable color types are truecolor or grayscale.');
end
end
catch
% if it is not an image but a variable
rightImage=rightImage;
end
% Find the size (columns and rows) of the left image and assign the rows to
% variable nrLeft, and columns to variable ncLeft
[nrLeft,ncLeft] = size(leftImage);
% Find the size (columns and rows) of the right image and assign the rows to
% variable nrRight, and columns to variable ncRight
[nrRight,ncRight] = size(rightImage);
% Check to see if both the left and right images have same number of rows
% and columns
if(nrLeft==nrRight && ncLeft==ncRight)
else
error('Both left and right images should have the same number of rows and columns');
end
% Check the size of window to see if it is an odd number.
if (mod(windowSize,2)==0)
error('The window size must be an odd number.');
end
% Check whether minimum disparity is less than the maximum disparity.
if (dispMin>dispMax)
error('Minimum Disparity must be less than the Maximum disparity.');
end
% Create an image of size nrLeft and ncLeft, fill it with zeros and assign
% it to variable dispMap
dispMap=zeros(nrLeft, ncLeft);
% Find out how many rows and columns are to the left/right/up/down of the
% central pixel based on the window size
win=(windowSize-1)/2;
tic; % Initialize the timer to calculate the time consumed.
for(i=1+win:1:nrLeft-win)
for(j=1+win+dispMax:1:ncLeft-win)
prevSAD = 65532;
temp=0.0;
bestMatchSoFar = dispMin;
for(dispRange=-dispMin:-1:-dispMax)
sad=0.0;
for(a=-win:1:win)
for(b=-win:1:win)
if (j-win+dispRange > 0)
temp=leftImage(i+a,j+b)-rightImage(i+a,j+b+dispRange);
if(temp<0.0)
temp=temp*-1.0;
end
sad=sad+temp;
end
end
end
if (prevSAD > sad)
prevSAD = sad;
bestMatchSoFar = dispRange;
end
end
dispMap(i,j) = -bestMatchSoFar;
end
end
% Stop the timer to calculate the time consumed.
timeTaken=toc;
MATLAB Code-SAD Right to Left Matching-Implementation 1
% *************************************************************************
% Title: Function-Compute Correlation between two images using the
% similarity measure of Sum of Absolute Differences (SAD) with Right Image
% as reference.
% Author: Siddhant Ahuja
% Created: May 2008
% Copyright Siddhant Ahuja, 2008
% Inputs: Left Image (var: leftImage), Right Image (var: rightImage),
% Window Size (var: windowSize), Minimum Disparity (dispMin), Maximum
% Disparity (dispMax)
% Outputs: Disparity Map (var: dispMap), Time taken (var: timeTaken)
% Example Usage of Function: [dispMap, timeTaken]=funcSADR2L('StereogramLeft.jpg', 'StereogramRight.jpg', 9, 0, 16);
% *************************************************************************
function [dispMap, timeTaken]=funcSADR2L(leftImage, rightImage, windowSize, dispMin, dispMax)
try
% Grab the image information (metadata) of left image using the function imfinfo
leftImageInfo=imfinfo(leftImage);
% Since SADR2L is applied on a grayscale image, determine if the
% input left image is already in grayscale or color
if(getfield(leftImageInfo,'ColorType')=='truecolor')
% Read an image using imread function, convert from RGB color space to
% grayscale using rgb2gray function and assign it to variable leftImage
leftImage=rgb2gray(imread(leftImage));
% Convert the image from uint8 to double
leftImage=double(leftImage);
else if(getfield(leftImageInfo,'ColorType')=='grayscale')
% If the image is already in grayscale, then just read it.
leftImage=imread(leftImage);
% Convert the image from uint8 to double
leftImage=double(leftImage);
else
error('The Color Type of Left Image is not acceptable. Acceptable color types are truecolor or grayscale.');
end
end
catch
% if it is not an image but a variable
leftImage=leftImage;
end
try
% Grab the image information (metadata) of right image using the function imfinfo
rightImageInfo=imfinfo(rightImage);
% Since SADR2L is applied on a grayscale image, determine if the
% input right image is already in grayscale or color
if(getfield(rightImageInfo,'ColorType')=='truecolor')
% Read an image using imread function, convert from RGB color space to
% grayscale using rgb2gray function and assign it to variable rightImage
rightImage=rgb2gray(imread(rightImage));
% Convert the image from uint8 to double
rightImage=double(rightImage);
else if(getfield(rightImageInfo,'ColorType')=='grayscale')
% If the image is already in grayscale, then just read it.
rightImage=imread(rightImage);
% Convert the image from uint8 to double
rightImage=double(rightImage);
else
error('The Color Type of Right Image is not acceptable. Acceptable color types are truecolor or grayscale.');
end
end
catch
% if it is not an image but a variable
rightImage=rightImage;
end
% Find the size (columns and rows) of the left image and assign the rows to
% variable nrLeft, and columns to variable ncLeft
[nrLeft,ncLeft] = size(leftImage);
% Find the size (columns and rows) of the right image and assign the rows to
% variable nrRight, and columns to variable ncRight
[nrRight,ncRight] = size(rightImage);
% Check to see if both the left and right images have same number of rows
% and columns
if(nrLeft==nrRight && ncLeft==ncRight)
else
error('Both left and right images should have the same number of rows and columns');
end
% Check the size of window to see if it is an odd number.
if (mod(windowSize,2)==0)
error('The window size must be an odd number.');
end
% Check whether minimum disparity is less than the maximum disparity.
if (dispMin>dispMax)
error('Minimum Disparity must be less than the Maximum disparity.');
end
% Create an image of size nrLeft and ncLeft, fill it with zeros and assign
% it to variable dispMap
dispMap=zeros(nrLeft, ncLeft);
% Find out how many rows and columns are to the left/right/up/down of the
% central pixel based on the window size
win=(windowSize-1)/2;
tic; % Initialize the timer to calculate the time consumed.
for(i=1+win:1:nrLeft-win)
for(j=1+win:1:ncLeft-win-dispMax)
prevSAD = 65532;
temp=0.0;
bestMatchSoFar = dispMin;
for(dispRange=dispMin:1:dispMax)
sad=0.0;
for(a=-win:1:win)
for(b=-win:1:win)
if (j+b+dispRange <= ncLeft)
temp=rightImage(i+a,j+b)-leftImage(i+a,j+b+dispRange);
if(temp<0.0)
temp=temp*-1.0;
end
sad=sad+temp;
end
end
end
if (prevSAD > sad)
prevSAD = sad;
bestMatchSoFar = dispRange;
end
end
dispMap(i,j) = bestMatchSoFar;
end
end
% Stop the timer to calculate the time consumed.
timeTaken=toc;
After running the above code (Implementation 1-Left to Right matching) on Tsukuba images (384×288 image resolution, disparity range [0,16], window size: 9×9), on my machine-Dell Studio XPS 435T, Intel Core i7-920 CPU, 2.67GHz, 9.00GB RAM, Windows Vista OS, MATLAB R2008a, i get an average run time of about 6.587 seconds.
Sample Image-Cones
