Preprocessing Correction for Micronucleus Image Detection Affected by Contemporaneous Alterations

This paper presents a method that detects and corrects alterations of 1) exposure; 2) out-of-focus; and 3) Gaussian noise affecting, contemporaneously, the images that are acquired in flow cytometer measurement devices. These alterations reduce image quality and interfere with correct micronucleus (MN) detection in a lymphocyte. The objectives of the proposed correction are given as follows: 1) to correctly process the image with the pattern-matching algorithm in order to detect the MN in human lymphocytes; 2) to minimize doubtful detections; and 3) to enhance the confidence that, in rejected images, MNs cannot be detected. Numerical and experimental tests confirm the validity of the proposed correction method and permit the evaluation of the upper and lower bounds of the admissible variation range of each alteration.