Elimination of blurred images of cellular fields to virtualize a slide

Abstract

The cell analysis, usually performed by specialists to diagnose a patient, has drawbacks. Since the sample has a short biological life cycle; moreover, it is prone to alterations by inappropriate handling. One solution to this problem is the digitalization and virtualizations of slides. The group of Development and Technological Innovation-DIT- of BIOSYS LTDA Company is developing a technique. This consists of slide digitization under a microscope (several shots of images of cell fields) and subsequent virtualization, while the user (Bacteriologist or technologist, etc.) performs their daily work. Due to blurring (caused by slide movement) at moment of reading the slide, not all pictures can be used for slide virtualization. For this reason, we must choose the image of each cell field, with less blur. This work presents the results of the implementation of a direct method based on the application of the Haar Wavelet Transform, to detect and eliminate blurred images which are not necessary in the creation of a virtual slide. The results were obtained from a database with 20 sets of cell fields, which for each field were available 15 images with different focus conditions.