crop_bbox#
- omnipose.utils.crop_bbox(mask, pad=10, iterations=3, im_pad=0, area_cutoff=0, max_dim=inf, get_biggest=False, binary=False)[source]#
Take a label matrix and return a list of bounding boxes identifying clusters of labels.
- Parameters
mask (matrix of integer labels) --
pad (amount of space in pixels to add around the label (does not extend beyond image edges, will shrink for consistency)) --
iterations (number of dilation iterations to merge labels separated by this number of pixel or less) --
im_pad (amount of space to subtract off the label matrix edges) --
area_cutoff (label clusters below this area in square pixels will be ignored) --
max_dim (if a cluster is above this cutoff, quit and return the original image bounding box) --
- Returns
slices
- Return type
list of bounding box slices with padding