A ROBUST DEEP LEARNING APPROACH FOR PRECISELY SEGMENTING CELLS IN MULTIPLEX TISSUE IMAGES
Daniel E. Winkowski, Jeni Caldara, Brit H Boehmer, Regan Baird · Journal of Pathology Informatics · 2022
negative tumor.Both the UNet model and HALO© were highly correlated with Pathologists' positive ER/PR scores.Because it could be adjusted on for each slide, HALO© had higher correlation with the consensus than UNet.These and similar computer-aided GT technologies can accelerate DL algorithm development by generating large-scale GT labels utilized in training and validation.