WBC differential count yields clinically relevant information about health and disease. Currently, pathologists manually annotate the WBCs (Differential WBC Count) according to their experience and intuition which is a subjective task and therefore not perfectly reproducible. Furthermore, this work is tedious and time consuming. This study aims at automation of DBC process so as to increase the productivity & eliminate human errors. The proposed system takes peripheral Leishman blood stain images as input and generates count of each WBC sub-type (DBC). There have been a few efforts in the past to automate the solution to this problem. In the opinion of the authors of this paper, the previous works on automation of differential WBC count have not laid emphasis on development of fast, parallel and scalable system. Such a fast system is needed as the number of DBC a pathology lab handles is enormous in number. This study focuses on…
View original post 347 more words
Manohar Kuse's Cyber 