Automated software system for halftone image recognition in Java

The relevance of the work is due to the growing need for automated analysis of biomedical images, such as cell micrographs. Manual analysis of such data is labor-intensive and prone to subjective errors. The aim of the study is to develop and test a software system in Java for the full cycle of halftone image processing: from noise suppression to object segmentation and measurement of their parameters. The work addresses the tasks of implementing and comparatively analyzing algorithms for noise suppression, contrast enhancement, binarization, and segmentation. The scientific novelty lies in the adaptation and integration of classical computer vision algorithms into a single pipeline based on Java, as well as in the application of a modified connected components algorithm using the Union-Find data structure to improve segmentation accuracy. The practical significance of the system is confirmed by its application to processing test cell images. As a result of the study, the quantitative parameters of the segmented fragments were obtained, such as area, linear dimensions, and integral brightness. The system demonstrates resilience to various types of noise and can be integrated into larger projects due to the use of the cross-platform Java language.