• VisWeek Home
• Contest Home
• Getting the Datasets
  • Acknowledgments & Citation
  • Dataset Content and Format
  • Synthetic data
  • Patient data
  • Utility Software
• Entering the Contest
• Driving Questions
• Submission Format
• Expert Evaluation
• Mailing List
• References & Learning more
• Prizes and Incentive
• Deadline and Notification

2009 contest has been deferred to 2010.

Driving Questions

Here are some general guestions that a neurosurgeon would want help answering when learning about the imaging data to plan the surgery. The questions revolve around the notion of "tract", which refers to a directionally coherent group of axons in the white matter. Tractography seeks to model axonal pathways by some algorithmic processing of Diffusion MRI.

• What is the relationship between the tracts near the tumor, and those in the rest of the brain? How large is the region of influence of the tumor on the rest of the tracts?
• What's the relationship between the tracts near the tumor, and the functional areas (as indicated by function MRI)?
• More generally, what is the relationship between tumor proximity and cortical connectivity?
• Given a particular set of tracts in the brain (perhaps interactively selected), do they run through the lesion, or to one or another side? How closely?
• Which are tracts that are infiltrated by the tumor, versus being displaced by it?
• Do you think there intact tracts that are masked by edema (high diffusivity)?

Submission Content and Format

The contest is open to everyone except contest organizers and judges. We invite submissions from individuals or teams, from industry or academia.

Our focus is visualization research, not the medical practice of surgical planning, so we do not want participants to try building fully-featured surgical planning applications, or to give advice on how these particular patients should be operated on. Rather, we encourage participants to demonstrate the prototype of a tool that supports a method or process (hopefully with an interactive component) for understanding this kind of data through analysis and visualization, with the general target of surgical planning.

To demonstrate their approach, participants will submit

• (required) A 2-page mini-paper in PDF format describing how the visualization and analysis techniques combined to help answer the driving questions.
• (required) Up to 12 additional images (beyond those in the paper) showing how the visualizations help answer the questions.
• (encouraged) An Mpeg, AVI, or Quicktime video (duration at most 10 minutes) showing the methods or processes in action. This will be most helpful for demonstrating the effectiveness of the proposed methods.

The 2-page write-up should not try to give any background information (e.g. about the imaging modality) that will the same for each contestant. We would like to learn the important information about how the analyses and visualizations were created, by what kind of software, and how long it took to generate the results.

We will not be running any software as part of the evaluations, but hope that the submitted images and video will illustrate how the contestant's methods help answer the driving questions above. Thus, we encourage participants to broaden their attention from just making a set of visualizations, to a dynamic process for using visualizations for exploration and planning.

Those planning to submit are strongly encouraged to subscribe to the mailing list so that you will receive updates and clarifications as they are sent out.

The URL for electronic submission will be announced later.

Expert Evaluation

• How long did it take to the results that are shown, both in terms of computation time and interactions with the user?
• How much parameter adjustment was required for this particular data, versus how automated is the tool?
• Are basic anatomical structures contextually indicated?
• What kinds of interactive exploration are supported? (e.g. Are Regions-of-Interest for tract seeding are interactively placed? What about tractography parameters?)
• Is significant learning or training required to understand the final rendered images?
• Do the methods require specialized hardware, or can it run on any modern PC with a decent graphics card?
• (optional) Do the final renderings include any indication of the uncertainty inherent in long-range tractography and tumor segmentation?
• (optional) How do these methods compare to prior published neurosurgical visualizations and exploration tools?

Mailing List

References & Learning more