Why are you trying to get the pixel value of a contour map? The pixel values of the contour map don't tell you anything about the image -- the shape of the contour does. What's worse is you're looking at the pixel values in a map that looks like it's been damaged by saving it as a JPG. This is meaningless information.
This is one attempt:
testpict = im2uint8(mat2gray(peaks(500)));
imshow(im2double(testpict)); hold on
The reason that the image is cast as double in the imshow() call is so that the pixel values are comparable to those used by contour(); otherwise, the contour lines won't be visible. Can the image be displayed in uint8 format with the contour overlay so that the datatip values are in uint8 range? I don't know. Maybe.
You can also just avoid the ambiguity of using a single-level contour map and just select a specific region and mask it off explicitly.
maskedpict = im2uint8(im2double(testpict).*m);
Either way, you can get the values of the image pixels using impixelinfo() or a datatip, as has been suggested already.