postprocessing options
ReconstructCellsIn3DFrom2DInstanceLabelsStrat
ReconstructCellsIn3DFrom2DInstanceLabelsStrat ()
If you’re analysing a z-stack dataset, features can overlap with features in other planes. By application of this strategy, findmycells tries to merge and assign the same label to them (“multi-match-traceback”). You need to run this step fo inspect features in 3D in the next step.
FillHolesStrat
FillHolesStrat ()
Specific conditions can lead to “holes” within the segmentations of your image features. This can make total sense, for instance if youre analyzing specifically the cells cytoplasm, and want to spare out the nucleus, or if youre analyzing ring-like features. In other cases, however, these “holes” may be artefacts and require correction. This strategy was designed for exactly this purpose: fill the “holes” in your segmented features, if they should not be there.
ApplyExclusionCriteriaStrat
ApplyExclusionCriteriaStrat ()
Another typical postprocessing step is to filter your data by applying exclusion criteria. Please check the following list of exclusion criteria with detailed descriptions to see which criteria are currently implemented in your installed findmycells version. Importantly, if you would like to use this strategy, please make sure to run this strategy as the last postprocessing strategy, to ensure that all other processing steps have been completed. - Minimum feature position relative to area ROI: If you provided ROIs that denote in which area of each image you’d like to quantify the image features, you can use this criterion to specify when to exclude detected features from the quantification. These are the options you can chose from, in order of increasing distance of your feature from the area ROI: within < intersects < touches < no overlap The option you select will always be the first relative position to be included. For example, if you select “intersects”, all features that are fully within your area ROI (classified as “within”) and all features that are intersected by the area ROI border (classified as “intersects”) will be kept, while all features that only touch the area ROI border (classified as “touches”) or that lie completely outside of it (classified as “no overlap”) will be excluded from further analyses. Note: if you are analyzing an image stack, a features` position can be classified differently depending on the image plane. Findmycells will always use the nearest classification. That means, if a feature was classified in one plane as “within” but only as “intersects” or even as “no overlap” in the other planes, the entire 3D feature will be classified as “within”. - Minimum feature size [px]: Every detected feature whose area is smaller than the specified pixel value will will be deleted from the segmentation masks. Note: if you are analyzing an image stack, this strategy will determine for each 3D feature the plane in which is has the largest area & then apply this exclusion criterion based on this area value. - Minimum planes covered (only relevant for image stacks): Similarly to “minimum feature size”, which checks for the expansion of your image features in x-y-dimensions, you can use this exclusion criterion to also check for a minimum size of your features in the z-dimension. The number you specify here will represent the minimum number of consecutive planes each image feature needs to cover to remain in your segmentation masks. For instance, if you specify it to be 3, all features that are only present in a single or in two consecutive planes will be deleted. Note: this exclusion criterion is, obviously, only relevant for projects that analyze image stacks. If you are working only with 2D images, this value will be ignored and 1 will be used as default.