Local Measurements in Recipe
Add local measurements as a step in the recipe! They can be displayed with a simple click after recipe loading and summary stats (e.g., mean, min, max, etc.) of the local measurement will appear in the global measurement report.
To add to the recipe, open Local Measure > setup and generate > customize formatting as desired > click “Add to Recipe” in upper-right corner. Full data tables of all local measurements can be exported per image as CSVs during a batch process, just as feature measurements can be today.
Dynamic Scale Factor
Setup your recipe to dynamically determine the scale factor based on the size of a recognized feature.
In Calibrate Scale, you can now setup a third mode which accepts a “sub-recipe” for detection and a defined dimension to measure. When the main recipe is run, this sub-recipe first detects the scale-determining feature and sets the scale factor according to the feature’s measured pixel and defined physical dimensions.
Dilation and Erosion in Physical Units
Uniform Dilation and Uniform Erosion depths can now be defined in physical units such as microns, rather than just pixels. Useful when defining custom regions-of-interest (e.g., border thickness) in physical units, expecially when the recipe needs to accomodate various magnifications.
Companion Area Fraction Measurements
Area fraction of “child” features, relative to their “parents” can now be made automatically as part of the “Companion Area” measurement under Measure Features.
To setup, store the “child” features in the Companion Image, and set the “parent” features as the layer to measure. Choose Measure Features > select layer > check “Companion Area” > add to recipe.
Improvements & Minor Changes
- Load large session from Batch Processor has more responsive loading spinner
- Improved file sorting in Batch Processor, Post Processor, and Deep Learning Trainer
- Up/down arrows in Post Processor and Deep Learning trainer can no longer produce mismatched reference and selection image sizes
- Load Companion Image now offers an option to force resize if necessary
- Local Measure will now open immediately when only one layer in the recipe
- Improved dark mode appearance in some preview windows
- Reject Features – Area Fraction (CRITICAL)
Features to be rejected using Area Fraction would measure incorrectly. The bug fix corrects the value in the Reject Features step to maintain detection prior to version 3.4.0.
1. Load Recipe
2. Edit the Reject Features step that uses Area Fraction
3. Change the value to the appropriate value to meet detection
4. Select ‘Accept’
5. Save Recipe
- Removing layers could cause an improper layer assignment to Feature Measurement steps
- MIPAR could fail to launch with certain outdated GPU drivers
- View > Manage Layers would cause Feature Measurement steps to reset to “All”
- GPU acceleration could cause freezing
- Load Companion Image would fail with a certain image type
- Intermittent issue with replacing reference images in Deep Learning Trainer
- Displayed scale factor units in some generated reports would not match the displayed numerical value
- Detection results from some deep learning model applications could appear offset from the reference image
- “Out of GPU memory” errors now caught in Adaptive Threshold
- Orthogonal Correlate would produce anomalous results in some cases
- Erroneous reference image replacement in Deep Learning Trainer and Post Processor will no longer break the Session file
- Reference image order could reset after first up/down move in Deep Learning Trainer
- Adding an inverse layer in Deep Learning Trainer would not always appear until the active image was changed
- In Post Processor an error could occur when opening a batch of results over the top of a previous batch
- Bug with Shift+clicking multiple images in Batch Processor
- Deleting a layer in Deep Learning Trainer or Post Processor would not enable the “Save Session” button
- Crop Image would not be applied to reference image in some cases upon loading recipe
- In Local Measure “Save with Colorbar” could fail
- In Local Measure “Intensity StdDev” measurement could fail
- Viewing mode dropdown no longer active in Simple Mode
- Global intensity measurements in Image Processor would be disabled for selection images that were converted to grayscale
- Summary stats of some local measurements would not report correctly from the IP and BP
- For some local measurements, CSV files from a batch process would contain the identical data for all images
- RTP did not output batch local measurements properly
- RTP could experience an issue and fail to run
- Headers in CSV files of batch local measurements had issue with units formatting
- “Output Local/Feature Meas” checkbox in BP and RTP would not re-enable after batch process finished
- “Auto” mode in Calibrate Scale could fail when search area extended beyond image frame
- DL models saved out of Apply Model could have issues being applied in future recipes
- Manual distance tool did not disable when using “Auto” mode
Need more help with this?
Chat with an expert now ››
Thanks for your feedback.