Skip to content

SlicinatorPlanning

Jump to bottom
Charl P. Botha edited this page Apr 28, 2016 · 1 revision

Introduction

The Slicinator is to be a DeVIDE viewer module for the slice-by-slice contouring of datasets.

An important difference with other solutions is that all contours are internally handled as distance fields (where possible) and masks otherwise. This greatly facilitates contour arithmetic: Adding / subtracting contours, interpolating between slices, etc.

A very basic skeleton has been started: http://code.google.com/p/devide/source/browse/trunk/devide/modules/viewers/Slicinator.py

General requirements

  • Input dataset (supplied by other DeVIDE modules) is shown in a combined sliceviewer (use CMSliceViewer for this).
  • Besides the sliceviewer, there is an orthogonal slice view (vtkImageColorViewer should help you here, see LarynxMeasurement.py for an example).
  • One can draw contours on the slice view. All drawn contours belong to the current default object.
  • At any time, a new object can be created, or one can switch to an old object. The selected object is the current context, and all new contours belong to the current object.
  • There are a number of per-slice tools available: draw, livewire, erase within newly drawn contour (done with subtraction), add newly drawn contour, region grow, smooth, etc. All of these internally operate on the pixel mask / distance field. Easy peasy.
  • There are a number of per-object tools available: interpolate, save, load, etc.
  • Drawn contours are also shown in the 3D combined sliceviewer view.
  • Also see the comments at the start of the current Slicinator for links to examples (for implementing line-drawing on slices in VTK etc) and also more hints as to the development.

Nice-to-have requirements

Other examples to study

Clone this wiki locally