Скачать или смотреть Binary Segmentation | Corrupted Images | Local Thresholding | Flat-Field Correction | python

Binary Segmentation of Images (Corrupted with a few Illumination patterns generated), using Local Thresholding mtehods

Otsu, Riddler-Calvard, Kittler-Illingworth algorithms with different block size parameter values

Niblack algorithm with different blocksize and weight parameter values (T(i,j) = μ(i,j) + w.σ(i,j))

Also with Global Thresholding with Flat-Field Correction to remove illumination patterns,

an implementation with python.

As can be seen, the global versions of the thresholding algorithms perform poorly on the illuminated (corrupted) images, the local versions had better results and when the flatfield filter coukd remove the illumination patterns to some extent, the global versions did a good job too.

(#python #imageprocessing #imageprocessingpython #segmentation #threshold )