Image Segmentation in Colorectal Tissue Slides Using Denoising Autoencoder

Colorectal cancer (CRC) is a  common tumor type with variable treatment course.  Given the high availability of histological slides and wealth of the prognostic information the slide images may provide, it is important to conduct corresponding image analysis in high-throughput fashion.   In this presentation, we will discuss a segmentation approach based on denoising autoencoder for colorectal whole slide images using annotated image patches.