A Fully Automated Approach to Segmentation of Irregularly Shaped Cellular Structures in EM Images

摘要: While there has been substantial progress in segmenting natural images, state-of-the-art methods that perform well such tasks unfortunately tend to underperform when confronted with the different challenges posed by electron microscope (EM) data. For example, EM imagery of neural tissue, numerous cells and subcellular structures appear within a single image, they exhibit irregular shapes cannot be easily modeled standard techniques, confusing textures clutter background. We propose fully automated approach handles these using sophisticated cues capture global shape texture information, learning specific appearance object boundaries. demonstrate our significantly outperforms techniques closely matches performance human annotators.