Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
  • Loading branch information
spichardo committed Jul 9, 2024
1 parent 6d106a8 commit a456b5b
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -75,7 +75,7 @@ doi: [10.1109/TUFFC.2023.3274046](https://doi.org/10.1109/TUFFC.2023.3274046). E

# Version log
- 0.3.5 - July 8thth, 2024
- New: Selection of mapping method for acoustic properties based on CT scanner model, kernel variant and resolution is now available when selecting the CT input in the first dialog. The selected scanner combination will be saved for future sessions. However, default values will be applied when changing the type from real CT, PETRA or ZTE. A reset button to restore default value is also available.
- New: Selection of mapping method for acoustic properties based on CT scanner model, kernel variant and resolution is now available when selecting the CT input in the first dialog. The selected scanner combination will be saved for future sessions. However, default values will be applied when changing the type from real CT, PETRA or ZTE. A reset button to restore default values is also available.
- New: Add advanced options for fine-tuning the processing. In this first iteration, these parameters are for the control of the domain generation in Step 1. A new button for "Advanced options" is now available.
- **Selection of Elastix Optimizer**. The default optimizer for Elastix, AdaptiveStochasticGradientDescent, works *almost* for every combination of CT->T1W and ZTE/PETRA->T1W. However, as much more testing has started to accumulate, we noticed that the CT->T1W was not working in some cases. We tested different options in Elastix, and the FiniteDifferenceGradientDescent and QuasiNewtonLBFGS seem to work well as a replacement. If your coregistration from CT to T1W is not working as expected, try one of these two other options. FiniteDifferenceGradientDescent seems to work better.
- **Control of fraction of trabecular**. By default, the percentage of trabecular is 80% using the line of sight of the trajectory. If the trajectory crosses very thin regions like the parietal bone, there may be a chance all region is considered trabecular, which may not be desired. A warning message will be printed out in the terminal output in that case. In such cases as the parietal bone, consider reducing a lower value of trabecular, such as 0.1 or 0.2.
Expand Down

0 comments on commit a456b5b

Please sign in to comment.