Boundary loss for highly unbalanced segmentation

Abstract

Widely used loss functions for CNN segmentation, e.g., Dice or cross-entropy, are based on integrals over the segmentation regions. Unfortunately, for highly unbalanced segmentations, such regional summations have values that differ by several orders of magnitude across classes, which affects training performance and stability. We propose a boundary loss, which takes the form of a distance metric on the space of contours, not regions. This can mitigate the difficulties of highly unbalanced problems because it uses integrals over the interface between regions instead of unbalanced integrals over the regions. Furthermore, a boundary loss complements regional information. Inspired by graph-based optimization techniques for computing active-contour flows, we express a non-symmetric $L_2$ distance on the space of contours as a regional integral, which avoids completely local differential computations involving contour points. This yields a boundary loss expressed with the regional softmax probability outputs of the network, which can be easily combined with standard regional losses and implemented with any existing deep network architecture for N-D segmentation. We report comprehensive evaluations and comparisons on different unbalanced problems, showing that our boundary loss can yield significant increases in performances while improving training stability. Our code is publicly available: https://github.com/LIVIAETS/surface-loss .

Runner-up for best paper award at MIDL 2019 [PMLR 102:285-296], invited for MedIA deep learning special issue (Volume 67, January 2021)

References in corpus (10)

• A Survey on Deep Learning in Medical Image Analysis
• Efficient Multi-Scale 3D CNN with Fully Connected CRF for Accurate Brain Lesion Segmentation
• Brain Tumor Segmentation with Deep Neural Networks
• Fully Convolutional Networks for Semantic Segmentation
• Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations
• Domain-adversarial neural networks to address the appearance variability of histopathology images
• AnatomyNet: Deep Learning for Fast and Fully Automated Whole-volume Segmentation of Head and Neck Anatomy
• 3D fully convolutional networks for subcortical segmentation in MRI: A large-scale study
• 3D Segmentation with Exponential Logarithmic Loss for Highly Unbalanced Object Sizes
• Hough-CNN: Deep Learning for Segmentation of Deep Brain Regions in MRI and Ultrasound

Cited by in corpus (27)

• ResUNet-a: a deep learning framework for semantic segmentation of remotely sensed data
• U-Net and its variants for medical image segmentation: theory and applications
• Medical Image Segmentation Using Deep Learning: A Survey
• Inter-slice Context Residual Learning for 3D Medical Image Segmentation
• Evaluation of machine learning algorithms for Health and Wellness applications: a tutorial
• Automatic Ischemic Stroke Lesion Segmentation from Computed Tomography Perfusion Images by Image Synthesis and Attention-Based Deep Neural Networks
• RDP-Net: Region Detail Preserving Network for Change Detection
• Do We Really Need Dice? The Hidden Region-Size Biases of Segmentation Losses
• A global method to identify trees outside of closed-canopy forests with medium-resolution satellite imagery
• Segmentation Loss Odyssey
• Distance Map Loss Penalty Term for Semantic Segmentation
• Geometric Loss for Deep Multiple Sclerosis lesion Segmentation
• Distribution-aware Margin Calibration for Semantic Segmentation in Images
• Joint Learning of Blind Super-Resolution and Crack Segmentation for Realistic Degraded Images
• An Elastic Interaction-Based Loss Function for Medical Image Segmentation
• Cutting-edge 3D Medical Image Segmentation Methods in 2020: Are Happy Families All Alike?
• DAM-AL: Dilated Attention Mechanism with Attention Loss for 3D Infant Brain Image Segmentation
• Spatially Covariant Lesion Segmentation
• Shape-aware Segmentation of the Placenta in BOLD Fetal MRI Time Series
• Topology-Aware Loss for Aorta and Great Vessel Segmentation in Computed Tomography Images
• On Mathews Correlation Coefficient and Improved Distance Map Loss for Automatic Glacier Calving Front Segmentation in SAR Imagery
• Anatomical and Diagnostic Bayesian Segmentation in Prostate MRI $-$Should Different Clinical Objectives Mandate Different Loss Functions?
• Shape Constrained CNN for Cardiac MR Segmentation with Simultaneous Prediction of Shape and Pose Parameters
• Bladder segmentation based on deep learning approaches: current limitations and lessons
• PGD-UNet: A Position-Guided Deformable Network for Simultaneous Segmentation of Organs and Tumors
• Residual Moment Loss for Medical Image Segmentation
• High-level Prior-based Loss Functions for Medical Image Segmentation: A Survey