Generation of Optimal Focused 2D and 3D Images for Cytopathological Analysis in Light Microscopic Systems

It is desirable for more realistic cytopathological analysis in light microscopic system to generate optimal focused 2D and 3D images, which help pathologists detect the locations and size of the cells in the specimen. Various studies using 2D multi-focus images in order to extract 2D focused images and 3D structure of the specimen have been actively applied in the literature for many years. In these studies, microscope platform is moved along the fixed range determined randomly on the Z axis for all types of specimens and magnification objectives. However, it is not guaranteed that significant information is extracted from the 2D multi-focus images and whole region of the specimen on the Z axis is scanned. Contrary to literature studies, our proposed study generates optimal focused 2D and 3D images for cytopathological analysis in the light microscopic systems by selecting optimal range and the number of 2D multi-focus images. Our approach is performed in two main stage; pre-process and Shape from Focus (SFF). In order to evaluate performance more realistic, two different non-reference quality metrics are used, and the success of proposed approach is proven by both quantitative and visual results.