Fig. 7: LACSS model trained with semi-programmably generated annotations.
From: Efficient end-to-end learning for cell segmentation with machine generated weak annotations
a Examples of training data and accompanying annotations generated semi-programmably. b Examples of model inferences. The left panel shows the input fluorescence image. The middle panels show the comparison of input image and the model predictions (yellow contours). The right panel shows the comparison of model prediction (white contour) with manual segmentation (color map). The bottom example represented a more challenging case with atypical image contrasts. Scale bars represent 20 micrometers.
