Recognition and Counting of WBCs using Wavelet Transform

Gholamreza Abedini, Mohammad Firouzmand, and Ali Razavi


WBC, Leukocyte, Recognition


In this work recognizing and counting White Blood Cells (WBC, leukocyte) were performed from microscopic images of peripheral blood smear particles in six levels. First, there were rate of congestion and noise in images and de-noising of color images used SURE-LET method which is based on wavelet transform. Second, it is preserving edges with usage of KUWAHARA filter. Third, edge detection with CANNY method is based on wavelet. Fourth, selecting the objects with attention to five types of leukocyte (MONOCYTEs, BASOPHILs, EOSINOPHILs, NEUTROPHILs and LYMPHOCYTEs) was based on relative object’s color channels. Fifth, there were separating selected objects using a few kinds of wavelet transforms, suitable filter bank and comparing them together, at the end, sixth, counting each one of them is based on last level and again using relative object’s color channels method. This work’s result is better than the previous methods, because amount of correctness was high.

Important Links:

Go Back