ImSimQAdform provides tools for both creating deformed DICOM image series on any modality, and also analysis tools for Deformable Image Fusion QA, which can be used to assess deformable and rigid image registration accuracy which may be part of general image fusion packages, or integrated within atlas-based contouring applications.

grid no def
grid def
Planning CT deformed to produce swollen anatomy to test image fusion software

Visualisation tools show fused image series, e.g. primary reference image versus transformed/deformed image that would ideally result in a perfect match. An example would be deformed images reconstructed as part of atlas-based segmentation process, where the ideal result would be the images reconstructed from the atlas library would be deformed to exactly match the new patient images, so that the contours would also exactly match the patient’s images. Any mismatches in the resulting deformed image would result in poorly fitting contours.

def im fusion qa visual
Original and deformed CT (generated from atlas segmentation s/w) overlaid for evaluation

Quantitative tools further report the Mutual Information Index and Correlation Coefficient, for both the whole 3D image data, and also for any selected volumes of interests derived from RT Structures.

quant table top
Report of slices and structures with pass / fail colour bars for Correlation Coefficient and Mutual Information Indices

This page was printed from OSL Web site at