Skip to content

Modified version of FingerCrop to solve the Histogram Equalization issue

Vedrana KRIVOKUCA requested to merge btp into master

Fixed the Issue #12 (closed) by modifying the HE function in to rescale the intensity of the histogram equalized image back to the range (0, 255), thereby avoiding the need to modify the binarization thresholds. Seems to work - "extracted" feature arrays are now binary and the score files do not contain "nan"

Closes #12 (closed)

Merge request reports