Analysis Strategies for smFISH Data Using Image Analysis Techniques

Nuclear image segmentation is one of the basic steps of quantitative computerised microscopy and digital pathology. Machine learning-based applications trained on sufficient supplemented images’ scale can be used to cope with problems associated with nuclear morphometry and nuclear density. Unfortunately, there has been little advancement due to inadequate diverse databases. To the point, star-convex shape-based segmentation approach incorporated in the library Stardist enables segmentation of densely packed or overlap of spherical nuclear structures in fluorescent images. It enhances specificity in the study of cellular architectures and allows the assessment of sub-cellular elements on higher resolution and accurate dimensions, enhancing studies of tumor heterogeneity and treatment response.