Synthesis of Annotated Colorectal Cancer Tissue Images from Gland LayoutDownload PDF

Anonymous

20 Jul 2021 (modified: 05 May 2023)MICCAI 2021 Workshop COMPAY Blind SubmissionReaders: Everyone
Keywords: Computational Pathology, Generative Adversarial Networks, Tissue Image Synthesis
Abstract: Generating annotated pairs of realistic tissue images along with their annotations is a challenging task in computational histopathology. Such synthetic images and their annotations can be useful in training and evaluation of algorithms in the domain of computational pathology. To address this, we present an interactive framework to generate pairs of realistic colorectal cancer histology images with corresponding tissue component masks from the input gland layout. The framework shows the ability to generate realistic qualitative tissue images preserving morphological characteristics including stroma, goblet cells and glandular lumen. We show the appearance of glands can be controlled by user inputs such as number of glands, their locations and sizes. We also validate the quality of generated annotated pair with help of the gland segmentation algorithm.
3 Replies

Loading