**Mentor** Stefan Atev, ViTAL Images, Inc.
- Qichuan Bai, The Pennsylvania State University
- Brittan Farmer, University of Michigan
- Eric Foxall, University of Victoria
- Xing (Margaret) Fu, Stanford University
- Sunnie Joshi, Texas A & M University
- Zhou Zhou, University of Michigan

**Project Description:**

Figure 1. Segmentation of the internal carotid artery (left).Vessel tree with the common, internal and external carotidarteries (right).

Accurate vessel segmentation is required in many clinicalapplications, such as identifying the degree of stenosis(narrowing) of a vessel to assess if blood flow to an organ issufficient, quantification of plaque buildup (to determine therisk of stroke, for example), and in detecting aneurisms whichpose severe risks if ruptured. Proximity to bone can posesegmentation challenges due to the similar appearance of boneand contrasted vessels in CT (Figure 1 – the internal carotidhas to cross the skull base); other challenges are posed by lowX-ray dose images, and pathology such as stenosis andcalcifications.

Figure 2. Cross section of vessel segmentation from CT data,shown with straightened centerline.A typical segmentation consists of a centerline that tracks thelength of the vessel, lumen surface and vessel wall surface.Since for performance reasons most clinical applications useonly local vessel models for detection, tracking andsegmentation, in the presence of noise the results can becomephysiologically unrealistic – for example in the figure above,the diameter of the lumen and wall cross-sections vary toorapidly.

Figure 3. Vessel represented as a centerline with periodicallysampled cross-sections in the planes orthogonal to thecenterline. Note that some planes intersect, which makes thisrepresentation problematic. The in-plane cross-sections of thevessel are shown on the right.The goal of this project is to design a method for refining avessel segmentation based on the following general approach:

Choose an appropriate geometric representations for vesselsegmentation (e.g., generalized cylinders) and derive theequations and methods necessary to manipulate it as requiredand to convert to and from the representation. One common, butsometimes problematic representation is shown in Figure 3.

Learn a geometric model for vessels based on therepresentation from a set of training data (for examplesegmentations obtained from low-noise clinical images). Examplemodel parameters:

- Relative rate of vessel diameter change as a function ofcenterline curvature

- Typical wall thickness as a function of lumen cross-section area

Learn an appearance model for the vessels that capturesdetails about how vessels appear in a clinical imaging modalitysuch as CT. For example:

- Radial lumen intensity profile in Hounsfeld units

- Rate of intensity change along the centerline

Compute the most likely vessel representation given a starting segmentation andthe learned geometric and appearance models.

The project will use real clinical data and many differenttypes of vessels.

**References:**

- C. Kirbas and F. Quek. “A review of vessel extractiontechniques and algorithms”. ACM Computing Surveys, vol. 36, pp.81–121, 2000.
- T. McInerney and D. Terzopoulos. “Deformable models inmedical image analysis: A survey”. Medical Image Analysis, vol.1, pp. 91 – 108, 1996.

**Prerequisites:**Optimization, Statistics and Estimation, Differential Equationsand Geometry. MATLAB programming.

**Keywords:**Vessel segmentation, shape statistics, appearance models