Fig. 1From: Application of convolutional neural networks towards nuclei segmentation in localization-based super-resolution fluorescence microscopy imagesComparison between conventional wide-field fluorescence microscopy image (left) and STORM super-resolution fluorescence microscopy image (right). Note the differences in nuclei structure between the STORM and wide-field images, with nuclei appearing as collections of discrete packages rather than as contiguous gradients. These textural differences result in incompatibilities when attempting to train a neural network for nuclei segmentation using one data set, then applying the trained network onto the other setBack to article page