Feature Extraction with Particles

From NAMIC Wiki
Jump to: navigation, search
Home < Feature Extraction with Particles

Key Investigators

Raul San Jose, James Ross (Brigham and Women's Hospital)

Project Description

Scale-space particles are a feature extraction tool that samples an image domain looking for second-order features based on Hessian information in both scale and space. Scale-space particles are implemented in Teem as part of the pull library.


  • Create a user-friendly interface to allow feature extraction based on space-scale particles in ROIs.
  • The primary goal is to develop a module that
    • can enable visual parameter exploration for whole image-domain feature extraction
    • can provide interactive features extraction in ROIs

Approach, Plan

  • Integrate the functionality of the pull library as a VTK class
  • Create a python module that relies on a vtkSlicerLogic class to orchestrate the VTK functionality.


  • Integration of latest Teem into Slicer (Special thanks to Hans Johnson and JC for their help).
  • VTK class is almost complete. The main standing issue it the creation of a stack of blur volumes in memory.