Detecting DAPI in a heterogeneous image set




I am having trouble finding a set of modules and parameters to accurately count DAPI signals in my image set. The main problem is it is an automatically acquired high-content set of images, and so the individual images are extremely heterogeneous in terms of number of cells and background, making thresholding without compromising counts very difficult. Has anyone used a similar image set and had success?




In general, this can be solved by

  • Setting upper and lower bounds on your threshold
  • Using the appropriate thresholding algorithm (RobustBackground is great when your nuclei are always sparse but isn’t the right algorithm to choose if they’ll be very dense in some images)
  • Doing background subtraction if necessary to remove heterogeneous background ahead of time.

Good luck!