May 24, 2006

From Utah Center for Neuroimage Analysis

Present: Tolga, Liz, Pavel, Samuel, Bryan, Chi-Bin, Melissa

Microneuroma data

Paul has demonstrated the final slice mosaics of microneuroma dataset.

Slice to slice affine registration on the microneuroma dataset is successful for roughly 75% of the slices. The issue of mismatched slices has to be addressed. Paul suspects the mismatches are due to contrast differences between the slices. The worst example of this can be seen in Section 10 of the microneuroma dataset, which has significantly worse contrast compared to Section 9 and Section 11.

Bryan insists on higher order transforms. Paul will try a higher order transform registration with variance minimization similar to the approach used in assembling the slices from tiles. Various approaches towards cascaded slice to slice registration were discussed, concerning the instability of cascading higher order transforms.

Axon tracking

Melissas trip to Germany to acquire higher resolution Zebrafish data was disappointing. The new acquisition method did not work for the Zebrafish tissue that we are interested in.

Liz said she will be working on the memory management issue in ZebraViewer next.

Misc

The next meeting will be June 28th at 10AM in SCI conference room.

(notes by Paul, please add anything I've missed).