Bell Jar 10.2.1

Changelog

• Fixed issue where file extension was not preserved on sharpen step
• Fixed issue with color images in cell detection

Bell Jar 10.2.0

Changelog

• Fixed issue where no Cuda device would be selected for cell detection

Bell Jar v10.1.0

Changelog

• Fixed the registration path to include rigid initial transform again
• Removed forced multithreading setting from ITK configuration to prevent a rare crash on some Windows systems
• Count now provides all regions regardless of the presence of counts for easy comparison across experiments

Note: if you are using Excel/Numbers, pivot tables can be used to sort and filter count data

Bell Jar v10.0.3

Changelog

• Added strict requirement for SAHI to prevent issue where some windows installs would revert to the oldest possible version breaking YoloV8 model support

Bell Jar v10.0.2

Changelog

• Fixed issue where zero length detections caused a crash
• Updated sample dataset with better quality images for detection and alignment
• Fixed a issue where a preprocessing step for alignment had been ommitted unintentionally

Bell Jar v10.0.1

Changelog

• Fixed issue with cell detection
• Adjustment to alignment parameters
• Changed timestamps to local time

Bell Jar v10.0.0

Changelog

• Added eccentricity cutoff for detection
• Added segmentation to isolate sections before alignment
• Improved speed of alignments

Bell Jar v9.5.2

Changelog

• Added area cutoff to cell detections
• Added nuclei model back
• New sharpen process, and max projection
• Sharpen and max projection now process on 16-bit images for increased quality

Bell Jar v9.5.1

Changelog

• Fixed issue where detections would not tile correctly
• Updated alignment parameters for better performance on non-server hardware

New

• Added sample dataset to release for testing, to use extract the contents, then click 'Open Guide' in Bell Jar and navigate to the sample section to get started

Bell Jar v9.5.0

Changelog

• Fixed issue causing alignments to shift downwards
• Updated detector model to better cover a variety of images
• Added option to change models in cell detection but only one model currently implemented
• Fixed issue where version was not displaying during load