Semantic segmentation of HeLa cells: An objective comparison between one traditional algorithm and four deep-learning architectures
Fig 6
Illustration of the image-label pairs created to train the U-Net architecture.
(a) A sample of regions, each 128 × 128 pixels placed next to each other as a montage. (b) Montage of the labels corresponding to the regions of (a). The labels contain four classes, from dark to bright: Nuclear Envelope, Nucleus, Cell, Background.