Fig. 6: Adaptability of the CNN.
From: A convolutional neural network segments yeast microscopy images with high accuracy
clb1-6Δ mutants were either excluded (left) or included (right) in the training set for the CNN, which was then tested on an image of clb1-6Δ cells from a later time point with even longer filaments, shown here. Note that the color of each cell is dependent on the internal numbering and therefore arbitrary. However, there are no fragmented filamentous cells on the right (green check marks) although there are segmentation errors when the strangely shaped cells are crowded (red arrow). Phase contrast image inverted for better visualization. Scale bar: 2 μm.
