v0.2.3 (20240115)

Changes since v0.2.1 (20220307):

• UI

  • Partial support of image colormap in UI (hard-coded colormap selection based on channel name).
  • Display plate map to make it easier to navigate in a plate when moving the stage.
  • Display some image quantification results in Data tab.

• Data Storage and Image Analysis

  • Use zstd compression in TIFF files.
  • Save TIFF images in a ZIP file.
  • Save image segmentation and quantification results in HDF5 file.
  • Image segmentation with U-Net for image analysis during acquisition.
  • Docker Compose file for setting up tensorflow serving server to serve the U-Net model.

• Python API

  • Set up samples and metadata.
  • Get NDImage and view with napari.
  • Initiate U-Net image segmentation from API.
  • Add hooks to allow deciding whether to acquire more images based on cell counts of acquired images.

• Configuration and Logging

  • Load channel config from config file.
  • Create a default user config file if no user config is found.
  • Log is written to files again in %LOCALAPPDATA%/NikonTiCtrl folder.

• Reliability Enhancement

  • Fix occasional json data corruption in index database.
  • Fix program crashing during acquisition due to frame data lost error reported by camera.
  • Retry by exposing the frame again to recover the multi-channel acquisition task after frame data lost error.
  • Workaround discovered
    • Showing the main window of the program (not minimizing) seems to greatly reduce the chance the frame data lost error during both live view and multi-channel acquisition.

v0.1.1 (20210719)

• Remove hard-coded channel/filter configurations and read config from files
• Compile version number into binary and write version number of the program in image files
• Output structured logs for elasticsearch and write logs into ProgramData folder
• Document dependencies and versions of the dependencies used during development and testing

v0.1.0 (20210614)

This is a reasonably stable version used in experiments since 6/14/2021.