PMOD Image Segmentation Tool Introduction (PSEG)

<< Click to Display Table of Contents >>

Navigation:  »No topics above this level«

PMOD Image Segmentation Tool Introduction (PSEG)

The PSEG tool implements a framework for image segmentation workflows, both for static and dynamic data. Currently, it offers solutions for different scenarios, namely:

1.PERCIST: Semi-automatic lesion segmentation and assessment of static data according to the PERCIST (PET Response Criteria in Solid Tumors) [1,2] methodology. As an additional feature, texture analysis can be applied within the detected lesions.

2.FUNCTIONAL (LOCAL MEANS): Semi-automatic segmentation of dynamic rodent PET studies into functional organs within only a few minutes.

3.CLUSTERING (K MEANS):  Automatic segmentation of dynamic data into clusters of "kinetically similar" pixels using the k-means algorithm.

4.CLUSTERING (Supervised):  Automatic segmentation of dynamic data into clusters of "kinetically similar" pixels corresponding to a set of prescribed time-activity curves (TACs).

5.MORPHOLOGICAL: Automatic segmentation of the input image data based on one of the selected general segmentation methods.

6.MACHINE LEARNING: Automatic segmentation of input image based on a Trained Network. It is mandatory to use input image with the same characteristics as the training images. This feature requires licensing of the PAI option.

All the Segmentation procedures, except PERCIST and FUNCTIONAL (LOCAL MEANS), return automatically the segments contours as VOIs.