Volume processing of fluorescent laser confocal microscopic cellular images

Abstract

The advents of light microscope techniques, together with advances in fluorescent labeling technologies, have revolutionized the study of cellular processes. Confocal microscopy has become a pivotal tool that has greatly impacted biological research at cellular and sub-cellular levels. It offers several advantages over conventional optical microscopy. However, several unique characteristics of confocal datasets pose serious challenges such as the noise and artifacts developed with confocal images and the intensity attenuation problem due to low axial resolution and light decay with depth. Currently, image processing, modeling, visualization and analysis on confocal microscopic datasets are, however, still done in 2D. In the case of traditional 2D sliceby-slice image processing strategy for the volumetric datasets, the image slices are analyzed only in the 2D plane where as the cross slice details (z axis) got ignored. On the other hand, volumetric image processing usually takes into consideration the interslice information and also preserves image structures across slices. Hence, instead of using 2D slice-by-slice strategy, treating confocal scanned volumes as a whole has its advantages. In this report, we provide a detailed literature study on various imageprocessing algorithms for confocal cellular image datasets. Our main goal of this research is to develop novel algorithms for 3D image processing methods for confocal cellular image enhancement, object feature extraction and subsequent volume visualization in an interactive and immersive VR environment.