Web-Based Manipulation of Multiresolution Micro-CT Images

Micro Computed-Tomography (μCT) scanning is opening a new world for medical researchers. Scientific data of several tens of gigabytes per image is created and usually requires storage on a common server such as Picture Archiving and Communication Systems (PACS). Previewing this data online in a meaningful way is an essential part of these systems. Radiologists who have been working with CT data for a long time are commonly looking at two-dimensional slices of 3D image stacks. Conventional web-viewers such as Google Maps and Deep Zoom use tiled multiresolution-images for faster display of large 2D data. In the medical area this approach is being adapted for high resolution 2D images. Solutions that include basic image processing still rely on browser external solutions and high-performance client-machines. In this paper we optimized and modified Brain Maps API to create an interactive orthogonal-sectioning image-viewer for medical μCT scans, based on JavaScript and HTML5. We show that tiling of images reduces the processing time by a factor of two. Different file formats are compared regarding their quality and time to display. As well a sample end-to-end application demonstrates the feasibility of this solution for custom made image acquisition systems.