Detection of poor quality peripheral blood smear images used in detection of leukocytes and erythrocytes

Close examination and precise analysis of the cells in blood smear images especially the Leukocytes and Erythrocytes is paramount for the accurate diagnosis of various diseases thereby aiding in the determination of the optimal route of treatment, for which the cells in the images have to be detected accurately using appropriate segmentation techniques. While acquiring the images for the study, certain images may turn out to be of poor quality, in which the cells may be over stained, clouded, coined or smudged. If these images are used in the segmentation process, the cells may not be properly segmented resulting in the degradation of the performance of the system. Hence there is a need to identify and prevent such poor quality images from entering the segmentation process. This paper proposes an automated method of detection of such poor quality images thereby preventing them from wrongly analyzed, thus increasing the specificity and sensitivity of the system by almost 20 percentages.