New hardware support

  • Olympus: U-LGPS LED and LDP light source
  • Teledyne Photometrics: Moment Global shutter scientific CMOS camera

New function/improvement

  • Improvement of TruAI (deep learning function)
    • The minimum image size for training files is 512×512.
    • In addition to VSI images, TIFF images can be used for training.
    • The number of image files used for training is displaied in the training setting
    • “Basic mode” and “Expert mode” for more detailed settings are selectable for the training.
    • Additional neural network types are available for the training.
    • Several types of training duration setting are available.
    • Check point can be set at any timing in the training setting.
    • Live display of inference results is possible and the results can be confirmed before the image acquisition.
    • Deep Learning inference and object counts can be run at once
    • Magic Wand is available for making annotations on the training images.
    • In case of Z-stack images processed, inference results are displayed for each frame.
    • Neural network processing information is showed on the properties of the processed images.
    • Additional GPUs of NVIDIA RTX3060Ti and RTX3070 are supported.
  • Restore function of cellSens condition to specified restore point is available (public information only)。
  • The data recovery process can be enabled or disabled. In case of “disable”, total time for the data acquisition can be reduced without the data recovery processing and this is beneficial especially for large volume data acquisition.
  • The period for the log collection can be specified.
  • Execution of the macro can be set in the Graphical Experiment Manager. The macro is executed just after the acquisitionon completed (specific processings only).
  • Additional ways of image export are available.
    • Raw data and Merged channels: The image of each channel and the merged image are exported.
    • Raw data and Merged channels with Burn in Info: The image of each channel and the merged image are exported, information is burned.
  • Improvement of NetCam function environment compatibility
  • Improvement of image cropping function
    • Acquired images can be cropped by specifying the XY position and size numerically.
    • ROI part can be saved as a new image.
  • CoolLED pE-300 white and pE-300 ultra LED light sources are supported in cellSens Standard