Study Of Microscopic Image Processing In Laser Confocal Scanning Microscopy System

Laser Confocal Scanning Microscopy (LCSM) is a microscopic instrument which provideshigh depth (Z-axis) resolution. It plays an important role in biomedical science. It can beused in studying morphology, recombination of three dimensional (3D) structures anddynamics of cell and tissue. It has been widely used for quantitative fluorescencemeasurement and quantitative imaging analysis in biomedical science. To take fulladvantage of LCSM, it is necessary to improve microscopic image processing technologyto develop a platform for LCSM to support high resolution image analysis in biomedicalapplications. To this end, this study investigated microscopic image restoration, imageprocessing and three dimensional (3D) image reconstruction from an array of 2D images inLCSM system. This study presents improved algorithms and new methods for LCSMimage processing in the following areas:1. The determination of Point Spread Function (PSF) is essential for image restoration.This paper has developed an improved deconvolution algorithm for estimating PSFusing error curve analysis. In addition, to improve the restoration of image detail inLCSM, the study improved NAS-RIF method to achieve the fine details restoration ofLCSM image.2. This paper has proposed and implemented anisotropic diffusion filtering based onadaptive filtering, accumulated image enhancing and spatial mixing image enhancingalgorithms for microscopic image processing, without losing of edge information andother details of image.3. Based on the study of two typical rendering methods for visualization of 3D data field,this paper proposes a new surface rendering algorithm of 3D image reconstructionbased on the examination of the surface information of object. Compared to traditionalsurface rendering method, the newly proposed algorithm simplifies surface rendering,which requires only height to generate vivid, realistic 3D dimension image. The 3Dimage reconstruction using volume rendering was discussed in detail. In LCSM 3Dimage reconstruction, Shear-Warp was incorporated into light reflection algorithm toimprove image processing speed. As a result 3D image was effectively reconstructedfrom a stack of 2D images.4. In addition to the above results, this study has constructed a framework and technicalplatform for biomedical image processing of LCSM. In conclusion, the improved biomedical images processing algorithms and platform weretested and verified in the LCSM image processing. The test results show that the improvedalgorithms and new methods as well as the platform can effectively process and analyzemicroscopic images of LCSM in practice. The results of this research can be a valuable forthe future research in biomedical image analysis.