After nuclear segmentation identifies all cells, classification determines which are marker-positive:
- Intensity thresholding: the mean signal in the marker channel is measured within each nucleus (or, for cytoplasmic or membrane markers, within a cell area estimated by expanding the nucleus), and cells above the threshold are positive
- Adaptive thresholding: the threshold varies by brain region to account for background differences
- Machine learning classifiers: a classifier trained on example cells combines several features (intensity, texture, size, context), which helps where intensity alone does not separate positive from negative cells
The choice of method depends on signal quality and the degree of background variation across the section.