Prior-aware Neural Network for Partially-Supervised Multi-Organ Segmentation

Abstract

Accurate multi-organ abdominal CT segmentation is essential to many clinical applications such as computer-aided intervention. As data annotation requires massive human labor from experienced radiologists, it is common that training data are partially labeled, e.g., pancreas datasets only have the pancreas labeled while leaving the rest marked as background. However, these background labels can be misleading in multi-organ segmentation since the "background" usually contains some other organs of interest. To address the background ambiguity in these partially-labeled datasets, we propose Prior-aware Neural Network (PaNN) via explicitly incorporating anatomical priors on abdominal organ sizes, guiding the training process with domain-specific knowledge. More specifically, PaNN assumes that the average organ size distributions in the abdomen should approximate their empirical distributions, a prior statistics obtained from the fully-labeled dataset. As our training objective is difficult to be directly optimized using stochastic gradient descent [20], we propose to reformulate it in a min-max form and optimize it via the stochastic primal-dual gradient algorithm. PaNN achieves state-of-the-art performance on the MICCAI2015 challenge "Multi-Atlas Labeling Beyond the Cranial Vault", a competition on organ segmentation in the abdomen. We report an average Dice score of 84.97%, surpassing the prior art by a large margin of 3.27%.

ICCV 2019

References in corpus (14)

• A Survey on Deep Learning in Medical Image Analysis
• Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
• Efficient Multi-Scale 3D CNN with Fully Connected CRF for Accurate Brain Lesion Segmentation
• Anatomically Constrained Neural Networks (ACNN): Application to Cardiac Image Enhancement and Segmentation
• A large annotated medical image dataset for the development and evaluation of segmentation algorithms
• Weakly- and Semi-Supervised Learning of a DCNN for Semantic Image Segmentation
• Constrained-CNN losses for weakly supervised segmentation
• Fully Convolutional Multi-Class Multiple Instance Learning
• Abdominal multi-organ segmentation with organ-attention networks and statistical fusion
• Deep Learning for Medical Image Segmentation
• DLTK: State of the Art Reference Implementations for Deep Learning on Medical Images
• Incorporating prior knowledge in medical image segmentation: a survey
• Semi-Supervised Multi-Organ Segmentation via Deep Multi-Planar Co-Training
• Recurrent Saliency Transformation Network: Incorporating Multi-Stage Visual Cues for Small Organ Segmentation

Cited by in corpus (18)

• Embracing Imperfect Datasets: A Review of Deep Learning Solutions for Medical Image Segmentation
• Asymmetric Non-local Neural Networks for Semantic Segmentation
• Label-set Loss Functions for Partial Supervision: Application to Fetal Brain 3D MRI Parcellation
• Learning Inductive Attention Guidance for Partially Supervised Pancreatic Ductal Adenocarcinoma Prediction
• Self-Supervision with Superpixels: Training Few-shot Medical Image Segmentation without Annotation
• AbdomenCT-1K: Is Abdominal Organ Segmentation A Solved Problem?
• DoDNet: Learning to segment multi-organ and tumors from multiple partially labeled datasets
• MS-KD: Multi-Organ Segmentation with Multiple Binary-Labeled Datasets
• Uncertainty-based graph convolutional networks for organ segmentation refinement
• Universal Lesion Detection by Learning from Multiple Heterogeneously Labeled Datasets
• Uncertainty-aware Incremental Learning for Multi-organ Segmentation
• Multi-Scale Attentional Network for Multi-Focal Segmentation of Active Bleed after Pelvic Fractures
• Discretely-constrained deep network for weakly supervised segmentation
• Multi-organ Segmentation via Co-training Weight-averaged Models from Few-organ Datasets
• Medical Image Segmentation with Limited Supervision: A Review of Deep Network Models
• ATSO: Asynchronous Teacher-Student Optimization for Semi-Supervised Medical Image Segmentation
• Context-aware virtual adversarial training for anatomically-plausible segmentation
• Atlas-aware ConvNetfor Accurate yet Robust Anatomical Segmentation