A convolutional neural network for common coordinate registration of high-resolution histology images

• DOI: 10.1093/bioinformatics/btab447
• Publication Date: 2021-06-14T19:12:33Z
• Authors: Aidan C Daly; Krzysztof J Geras; Richard Bonneau
• Keywords: Image registration; Transfer of learning

Registration of histology images from multiple sources is a pressing problem in large-scale studies spatial -omics data. Researchers often perform 'common coordinate registration', akin to segmentation, which samples are partitioned based on tissue type allow for quantitative comparison similar regions across samples. Accuracy such registration requires both high image resolution and global awareness, mark difficult balancing act contemporary deep learning architectures.