Biopolymer segmentation from CLSM microscopy images using a convolutional neural network

Confocal microscopy allows visualization of biopolymer networks at the nano scale. Analyzing structure and assembly protein from images requires a segmentation process. This has proven to be challenging due multiple possible sources noise in as well exhibition out-of-focus planes. Here, we present deep learning-based procedure for confocal laser scanning networks. Utilizing an encoder-decoder network architecture, our neural achieved dice score 0.88 segmenting filamentous temperature sensitive Z proteins chloroplasts Physcomitrella patens, moss.