{"id":4556,"date":"2013-05-10T09:00:49","date_gmt":"2013-05-10T13:00:49","guid":{"rendered":"https:\/\/blogs.mathworks.com\/pick\/?p=4556"},"modified":"2013-05-09T13:58:06","modified_gmt":"2013-05-09T17:58:06","slug":"colorspaces-out-the-wazoo","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/pick\/2013\/05\/10\/colorspaces-out-the-wazoo\/","title":{"rendered":"Colorspaces out the Wazoo!"},"content":{"rendered":"\r\n

Brett<\/a>'s Pick this week is \"Colorspace Transformations\"<\/a>, by Pascal Getreuer<\/a>.<\/p>

For inspiration for this week's Pick, I went back to my post<\/a> from April of last year, in which I asked readers to suggest files to feature. I started to write up Pascal's submission long ago, but somehow got sidetracked, and I wanted to circle back around to it.<\/p>

File Exchange champ Oliver Woodford<\/a> casually suggested \"a couple of useful ones.\" (His other selection was featured last August<\/a> by Jiro.) Both of his \"Picks\" are quite useful.<\/p>

So, what's so great about this colorspace converter? In particular, I really like the fact that Pascal's code extends the colorspace conversion capabilities beyond those offered by the Image Processing Toolbox, and that Pascal has included C-code (and instructions for calling it via the MEX interface<\/a>) to make them faster.<\/p>

Colorspace Transformations facilitates conversion to any of these colorspaces:<\/p>

'RGB'              sRGB IEC 61966-2-1\r\n'YCbCr'            Luma + Chroma (\"digitized\" version of Y'PbPr)\r\n'JPEG-YCbCr'       Luma + Chroma space used in JFIF JPEG\r\n'YDbDr'            SECAM Y'DbDr Luma + Chroma\r\n'YPbPr'            Luma (ITU-R BT.601) + Chroma\r\n'YUV'              NTSC PAL Y'UV Luma + Chroma\r\n'YIQ'              NTSC Y'IQ Luma + Chroma\r\n'HSV' or 'HSB'     Hue Saturation Value\/Brightness\r\n'HSL' or 'HLS'     Hue Saturation Luminance\r\n'HSI'              Hue Saturation Intensity\r\n'XYZ'              CIE 1931 XYZ\r\n'Lab'              CIE 1976 L*a*b* (CIELAB)\r\n'Luv'              CIE L*u*v* (CIELUV)\r\n'LCH'              CIE L*C*H* (CIELCH)\r\n'CAT02 LMS'        CIE CAT02 LMS<\/pre>
That's a lot of different ways to represent a color image!<\/p>
Whenever I am asked to segment a color image--and trust me, that's pretty often!--my initial thought is to look at the R,G, and B colorplanes individually to understand where the information I'm after lies. I do this so often that I wrote and shared a utility<\/a> that facilitates this exploration process. (In \"Advanced Mode,\" ExploreRGB shows--in addition to RGB [and R, G, and B]--HSV, YCbCr, and L*a*b* versions of the image. I can often find a single-plane version of the original image in one of these transformed colorspaces that makes the segmentation easy. For instance, if I needed to isolate the central yellow pepper in the \"peppers.png\" image that ships with the Image Processing Toolbox:<\/p>
\"\" <\/p>
I might consider (ironically) starting from the blue chrominance image (i.e., the second colorplane of the YCbCr representation of the image):<\/p>
\"\" <\/p>
Starting with that image plane:<\/p>
mask = ~im2bw(img,0.24);\r\ncc = bwconncomp(mask);\r\nstats = regionprops(cc,'Area'<\/span>);\r\nA = [stats.Area];\r\n[~,biggest] = max(A);\r\nmask(labelmatrix(cc)~=biggest) = 0;\r\nmask = imfill(mask,'holes'<\/span>);\r\nimshow('peppers.png'<\/span>)\r\nshowMaskAsOverlay(0.7,mask,'c'<\/span>);\r\n<\/pre>
\"\" <\/p>
Working in a non-standard colorspace made this segmentation problem easier than it might otherwise have been, and certainly easier than doing it in RGB-space would be.<\/p>
(showMaskAsOverlay<\/tt><\/a> is a utility I shared for overlaying a transparent mask on an image.)<\/p>
Okay, so back to Pascal's file.<\/p>
With a little bit of code, I can use \"Colorspace Transforms\" to examine quickly many additional representations of the original image:<\/p>
cs = {'RGB'<\/span>,'YCbCr'<\/span>,'JPEG-YCbCr'<\/span>,'YDbDr'<\/span>,'YPbPr'<\/span>,'YUV'<\/span>,...<\/span>\r\n'YIQ'<\/span>,'HSV'<\/span>,'HSL'<\/span>,'HSI'<\/span>,...<\/span>\r\n'XYZ'<\/span>,'Lab'<\/span>,'Luv'<\/span>,'LCH'<\/span>,'CAT02 LMS'<\/span>};\r\npeppers = imread('peppers.png'<\/span>);\r\nfigure('color'<\/span>,'w'<\/span>);\r\nax = tight_subplot(numel(cs),4);\r\nind = 1;\r\nfor<\/span> ii = 1:numel(cs)\r\n    newim = colorspace(['->'<\/span> cs{ii}],peppers);\r\n    if<\/span> max(newim(:)) > 1\r\n        newim = newim\/max(newim(:));\r\n    end<\/span>\r\n    axes(ax(ind));\r\n    imshow(newim,[]);title(cs{ii});\r\n    for<\/span> jj = 1:3\r\n        ind = ind+1;\r\n        axes(ax(ind));\r\n        imshow(newim(:,:,jj),[]);title(sprintf('%s(:,:,%i)'<\/span>,cs{ii},jj));\r\n    end<\/span>\r\n    ind = ind+1;\r\nend<\/span>\r\nexpandAxes(ax)\r\n<\/pre>
\"\" <\/p>
Now I have a lot more options at my disposal for segmenting images!<\/p>
P.S. I know you can't really see what's happening in those tiny thumbnails*, but then that's the beauty of expandAxes<\/tt><\/a>; if you were to run the code snippet above, you would be able to click on any of those tiny axes and expand them to full-frame. And you could right-click on any expanded axes to export the image to the Workspace!<\/p>
*P.P.S. If your image is large, doing what I just did could be a very bad idea. These aren't really \"thumbnails\"; even if the visualization is small, each of those image objects has in its \"cdata\"<\/a> container a full copy of the image it reflects.<\/p>
Thank you, Oliver, for the suggestion. And thank you<\/i>, Pascal, for the very useful submission. Swag on the way to both of you!<\/p>
As always, I welcome your thoughts and comments<\/a>. Or leave feedback for Pascal here<\/a>.<\/p>
(tight_subplot<\/tt><\/a> was a previous Pick-of-the-Week.)<\/p>