FluoRender: Visualization-Based and Interactive Analysis for Multichannel Microscopy Data

点击翻译按钮获取中文摘要

Summary FluoRender is a software package for visualizing and analyzing 3D and 4D (3D over time) fluorescence microscopy data. This project will serve the needs of biologists utilizing confocal microscopy for understanding cell development in many organisms and addresses the big-data problem from the massive increase of imaging data from modern high-resolution fluorescence microscopes. Specific Aim 1: Visualization of an extended number of volume channels: FluoRender will be enhanced with the multichannel visualization capability by simultaneously supporting several tens to hundreds of channels, which can be acquired from multispectral imaging devices or by registering data of multiple scans. FluoRender will take advantage of the latest volume rendering techniques to visualize significantly improved signal intensity detail compared to pseudo-surfaces. Specific Aim 2: Interactive comparison and organization of volume channels: A package of measures will be implemented in FluoRender for directly comparing volume channels. Leveraging the OpenCL programming interface, shape comparisons will be performed interactively on graphics hardware, allowing compound measures for complex morphology as well as immediate visual feedback via multichannel visualization. Interactive comparison will further enable the development of functions for semiautomatic channel organization and multichannel colocalization analysis. Specific Aim 3: 4D tracking of structures with irregular and changing shapes: Tracking irregularly shaped and shape-changing structures will substantially expand FluoRender's application for developmental and morphological studies of intracellular organelles, cells, and tissues. This will include a comprehensive tracking system that integrates different modules and allows them to work in an iterative and integrated environment, allowing user-guided, progressive refinement of the segmentation and tracking results. Specific Aim 4. Fully hardware-accelerated and customizable computing modules: FluoRender will be restructures using compute modules based on the OpenCL standard, which provides not only hardware-accelerated execution speed, but also convenience for customization and reuse. Computing modules will be integrated with visualization features, enabling interactive and visualization-centered analysis. Users will also be able to reorganize and build modules to customize specific workflows for great adaptability.