National Institute of Standards and Technology

Interactivity over 1000 z-stacks


Scientific studies of 3D morphology over a very large number of shapes face a challenge of viewing and comparing thousands of shapes. The goal of our work is to create a web application that can support visual inspections in a browser to enable collaborative discoveries.

This work is motivated by a biological study about stem cell and scaffold interactions. To derive significant conclusions about these cells, biologists must be able to view and study a large number of 3D images rapidly. The problem addressed is to design a web application that enables viewing high numbers of 3D data sets interactively while managing memory usage.

Figure 1 illustrates a typical visual inspection sequence that involves changing spatial resolutions and the number of cells being scrutinized. The interactive system is available at the URL.

Visualization of 3D image segmentation results.
Visual inspection of 3D image segmentation includes (a) viewing multiple cells, (b) bringing up high resolution version for analysis, zooming and rotation, and (c) sorting and ranking all cells based on their 3D shape characteristics.


We approached the problem by creating the Mass Stem Cell 3D Web Viewer, which is comprised of two components:

  • a pre-processing computational pipeline that creates a pyramid representation of segmented 3D cells allowing the Viewer to be scalable and
  • a web-based JavaScript application supporting interactive 3D inspection functions.
  • The Viewer supports scientists with the following features:

  • Selecting cells from multiple scaffolds
  • Translating, rotating and zooming
  • Color-coding and sorting based on file name, volume, surface, depth, and three principal axes lengths
  • Tooltip display with scaffold, file name and pyramid resolution for each cell
  • Downloading a csv file with all 3D shape properties
  • Help and instructions
  • We have prototyped a web solution that enables viewing 1147 cells acquired on 10 different scaffolds (approximately 110 cells per scaffold, 122,949 images and 122 GB of data). The Viewer runs on client computers with minimal RAM specifications. The key contributions of our work lie in:

  • Enabling scalable transmission and viewing of thousands of 3D data sets
  • Providing interactive visual inspection functions
  • Analyzing the Viewer’s client-side RAM and server-side disk storage requirements

  • Lead Organizational Unit:



    ITL-Software and Systems Division
    Information Systems Group


    Peter Bajcsy
    Phone: 301.975.2958

    Date created: April 10, 2014 | Last updated: