A Statistical Focusing Metric for Fluorescent Microscopy

Abstract

Autofocusing has critical importance for imaging systems with motorized microscopes in healthcare. It is not possible to diagnose on a picture that captured out of focus. In the literature, it has been observed that there are limited studies on focusing methods specific to fluorescent microscopes. In this study, a focus metric is proposed special to fluorescent microscope images. The proposed metric evaluates the responses generated by gradient filters of varying kernel size at a pixel point, and takes into account their standard deviations. The proposed method was tested on lung and breast tissue samples obtained with fluorescent microscope, and experimental results were reported. It is shown that the developed method overperforms the local gradient filters and produces an average error of 0.16 levels.