{"id":2660,"date":"2017-07-24T15:57:47","date_gmt":"2017-07-24T19:57:47","guid":{"rendered":"https:\/\/blogs.mathworks.com\/steve\/?p=2660"},"modified":"2022-08-24T17:01:29","modified_gmt":"2022-08-24T21:01:29","slug":"colormap-test-image","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/steve\/2017\/07\/24\/colormap-test-image\/","title":{"rendered":"Colormap Test Image"},"content":{"rendered":"

Today I want to tell you how and why I made these images:<\/p>

\"\" <\/p>

After the MATLAB R2014b release, I wrote several blog posts (Part 1<\/a>, Part 2<\/a>, Part 3<\/a>, and Part 4<\/a>) about the new default colormap, parula<\/a>, that was introduced in that release.<\/p>

Sometime later, I came across some material by Peter Kovesi about designing perceptually uniform colormaps<\/a> (or colourmaps, as Peter writes it).<\/p>

I was particularly intrigued by a test image that Peter designed for the purpose of visually evaluating the perceptual characteristics of a colormap.<\/p>

The image is constructed by superimposing a sinusoid on a linear ramp, with the amplitude of the sinusoid getting smaller as you move away from the top row. Here are three cross-sections of the image: row 1, row 64, and row 128.<\/p>

url = 'http:\/\/peterkovesi.com\/projects\/colourmaps\/cm_MATLAB_gray256.png'<\/span>;\r\nI = im2double(imread(url));\r\n\r\nsubplot(3,1,1)\r\nplot(I(1,:))\r\naxis([1 512 0 1])\r\ntitle('Row 1'<\/span>)\r\n\r\nsubplot(3,1,2)\r\nplot(I(64,:))\r\naxis([1 512 0 1])\r\ntitle('Row 64'<\/span>)\r\n\r\nsubplot(3,1,3)\r\nplot(I(128,:))\r\naxis([1 512 0 1])\r\ntitle('Row 128'<\/span>)\r\n<\/pre>\"\" 
Here is the basic code for making this test image. I'm going to vary Kovesi's image slightly. I'll add an extra half-cycle of the sinusoid so that it reaches a peak at the right side of the image, and I'll add a full-range linear ramp section at the bottom. (If you look closely at the cross-section curves above, you'll see that the linear ramp goes from 5% to 95% of the range.)<\/p>
First, compute the ramp.<\/p>
num_cycles = 64.5;\r\npixels_per_cycle = 8;\r\nA = 0.05;\r\nwidth = pixels_per_cycle * num_cycles + 1;\r\nheight = round((width - 1) \/ 4);\r\nramp = linspace(A, 1-A, width);\r\n<\/pre>
Next, compute the sinusoid.<\/p>
k = 0:(width-1);\r\nx = -A*cos((2*pi\/pixels_per_cycle) * k);\r\n<\/pre>
Now, vary the amplitude of the sinusoid with the square of the distance from the bottom of the image.<\/p>
q = 0:(height-1);\r\ny = ((height - q) \/ (height - 1)).^2;\r\nI1 = (y') .* x;\r\n<\/pre>
Superimpose the sinusoid on the ramp.<\/p>
I = I1 + ramp;\r\n<\/pre>
Finally, add a full-range linear ramp section to the bottom of the image.<\/p>
I = [I ; repmat(linspace(0,1,width), round(height\/4), 1)];\r\n\r\nclf\r\nimshow(I)\r\ntitle('Colormap test image'<\/span>)\r\n<\/pre>\"\" 
Last week, I posted Colormap Test Image<\/a> to the File Exchange. It contains the function colormapTestImage<\/tt>, which does all this for you.<\/p>
I = colormapTestImage;\r\n<\/pre>
The function has another syntax, too. If you pass it the name of a colormap, it will display the test image using that colormap. For example, here is the test image with the old MATLAB default colormap, jet<\/tt>.<\/p>
colormapTestImage('jet'<\/span>)\r\n<\/pre>\"\" 
This test image illustrates why we replaced jet<\/tt> as the default MATLAB colormap. I have annotated the image below to show some of the issues.<\/p>
\"\" <\/p>
Now compare with the new default colormap, parula<\/tt>.<\/p>
colormapTestImage('parula'<\/span>)\r\n<\/pre>\"\" 
I think that illustrates what we were trying to achieve with parula<\/tt>: perceptual fidelity to the data.<\/p>
Since I'm talking about parula<\/tt>, I'll finish by mentioning that we need some very subtle tweaks to parula<\/tt> in the R2017b release. So you can compare, I'll show you the original version that shipped with R2014b.<\/p>
colormapTestImage('parula_original'<\/span>)\r\n<\/pre>\"\" 
Readers, can you tell what is different? Let us know in the comments.<\/p>