文件列表:

comp_color_win32 (0, 2002-03-13)
comp_color_win32\cygwin1.dll (768840, 2002-01-21)
comp_color_win32\difCalc.exe (571904, 2002-03-14)
comp_color_win32\lena.ppm (607545, 2002-03-11)
comp_color_win32\lena2.ppm (607545, 2002-03-11)
comp_color_win32\sample.ngb (233, 2002-03-13)

Let me introduce what is difCalc, in which situation you could use it, and in which situation it will not help you. DifCalc is tool that computes objective differences between a reference image and a second image to compare. (please note that those images should be in binary PPM format) As it computes both a quality value and an image of differences, it allows you to really see in which area of the picture the differences are mostly located. As it computes some objective quality values, you do not rely on any human interpretation, and so the measurements are totally reproducible. The range of processing that could be applied to an image is quite large, and so a single difference value (like as an example PSNR) is often insufficient. As difCalc allows you to select a whole range of difference descriptors, you can really focus your quality evaluation on a specific part. Unfortunately it also means that using difCalc you will not have a single value that you can directly use, but a set of values (one for luminance, one for chrominance,...). It is up to you to determine the relative importance of each of the criterions in order to obtain a single quality value. This will require a little more work than a PSNR value, but also gives you a lot of flexibility. As the criterions are objectives, they do not take into account human preferences. It means that for a perceptual image processing, at the end you will still have to check the results with human observations.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：