Combalia M, Pérez-Anker J, García-Herrera A, Alos L, Vilaplana V, Marques F, et al.. Digitally Stained Confocal Microscopy through Deep Learning. In International Conference on Medical Imaging with Deep Learning (MIDL 2019). In Press.

Abstract

Specialists have used confocal microscopy in the ex-vivo modality to identify tumors with an overall sensitivity of 96.6% and specicity of 89.2%. However, this technology hasn't established yet in the standard clinical practice because most pathologists lack the knowledge to interpret its output. In this paper we propose a combination of deep learning and computer vision techniques to digitally stain confocal microscopy images into H&E-like slides, enabling pathologists to interpret these images without specic training. We use a fully convolutional neural network with a multiplicative residual connection to denoise the confocal microscopy images, and then stain them using a Cycle Consistency Generative Adversarial Network.