Efficient, learning based computational methods:

• Active Models
• Marginal Space Learning

Hierarchical, part based statistical modeling using machine learning for:

• Object detection and segmentation
• Integrating detection and tracking

Applications to medical imaging:

• 2D/3D detection or structures of interest: tumors, nodules, landmarks, organs
• 2D/3D thin, flexible structure localization and tracking (e.g catheter, guidewire)
• 2D/3D vessel segmentation
• 3D segmentation of brain structures
• 3D organ segmentation (liver, kidney, lungs, polyps, etc)