Motion Recovery and Segmentation of Cell Image in Biomedical Microscopy
Through quantitative research at the molecular level, explore the pathological mechanism of diseases, and reveal the origin and formation of life; it is the main development idea of biomedical field at present and in the future. From qualitative analysis to quantitative analysis of ultrastructure, it has gradually become the research direction of biomedicine, which also means the method and means of quantitative processing of electron microscope images are needed. Motion restoration is an image registration problem in image processing algorithm specific application in time series images, and image registration and image segmentation are two widely used and interrelated directions in the field of image processing and analysis. Therefore, to study the problem of image processing at the level of cell resolution, to make it a fully automatic, human-free, efficient and accurate computer-aided technology, it is of great significance to the field of biomedicine. In order to improve segmentation efficiency, the rapidity of image segmentation is studied. In this paper, genetic algorithm with better optimization performance is used to optimize multiple thresholds; experimental results show that this method can find a group of optimal solutions more accurately. It is much less time consuming than simulated annealing algorithm (SA) and exhaustive method. Meanwhile, compared with Otsu and maximum entropy method, the segmentation method combined with genetic algorithm takes slightly more time. However, higher quality image segmentation results can be obtained. The compromise result is that the segmentation method combined with genetic algorithm can achieve better results.