Involving logical clinical knowledge into deep neural networks to improve bladder tumor segmentation

Xiaodong Yue, Xiao Huang, Zhikang Xu, Yufei Chen, Chuanliang Xu

Abstract: Highlights•Formulate clinical logic rules of bladder tumor localization.•Utilize Graph Convolutional Network (GCN) to embed the clinical logic rules into the latent feature space.•Propose a novel bladder tumor segmentation method through involving clinical logic rules into DCNNs.•Reconstruct the loss objective of Deep Convolutional Neural Networks (DCNNs) to harness the tumor segmentation consistent with both image annotation and logic rules.•The proposed method is much less dependent on data and can achieve precise segmentation on limited number of images.