Analysis: Evaluate performance of rigid brain motion correction on three tracer datasets

ENHANCE-PET / FALCON

FALCON is a Python-based software application designed to facilitate PET motion correction, both for head and total-body scans. Our program is built around the fast 'greedy' registration toolkit, which serves as the registration engine. With FALCON, users can enjoy a streamlined experience for implementing motion correction.

GNU General Public License v3.0

36 stars 6 forks source link

Analysis: Evaluate performance of rigid brain motion correction on three tracer datasets #37

Closed Keyn34 closed 1 year ago

Keyn34 commented 2 years ago

Now that the ZUBAL case is sorted, evaluation of FALCON's capabilities concerning rigid brain motion correction will be assessed based on three tracer datasets (AMT, FDG, FET) with dice score evaluation.

[x] Collect ~10 datasets for each tracer
[x] extract full head mask for evaluation purposes
[x] tweak/prepare code to let FALCON run on all datasets
[x] gather dice scores before & after registration and visualize the metric

LalithShiyam commented 2 years ago

@Keyn34: Please let me know the tracers (along with their scanners), before you start the analysis!

Keyn34 commented 2 years ago

Update: adjusted deadline due to delays (thanks, Ubuntu...) Currently scanning through FET data.

Keyn34 commented 2 years ago

25 datasets (5 AMT, 10 FET and 10 FDG [5 Siemens, 5 Philips]) were collected and delineated for evaluation purposes.

We will perform the evaluation soon.

Keyn34 commented 2 years ago

Redoing the segmentations for FDG now: Siemens starting frame: 25 Philips starting frame: 8

Additional ToDos:

[ ] get framing for all datasets
[x] rerun evaluation pipeline on FDG set after segmentation individually

Tasks will be probably finished by Monday next week (11.07.2022).

Keyn34 commented 1 year ago

Analysis was done so far, waiting for Philips framing.

Philips segmentations were done with anisotropic filtering to enhance the segmentation quality.