Semi-supervised Medical Image Segmentation through Dual-task Consistency (2020-09-09T00:00:00.000000Z)

TL;DR

This work proposes a novel dual-task-consistency semi-supervised framework that can largely improve the performance by incorporating the unlabeled data and outperforms the state-of-the-art semi- supervised learning methods.

Abstract

Deep learning-based semi-supervised learning (SSL) algorithms have led to promising results in medical images segmentation and can alleviate doctors' expensive annotations by leveraging unlabeled data. However, most of the existing SSL algorithms in literature tend to regularize the model training by perturbing networks and/or data. Observing that multi/dual-task learning attends to various levels of information which have inherent prediction perturbation, we ask the question in this work: can we explicitly build task-level regularization rather than implicitly constructing networks- and/or data-level perturbation and then regularization for SSL? To answer this question, we propose a novel dual-task-consistency semi-supervised framework for the first time. Concretely, we use a dual-task deep network that jointly predicts a pixel-wise segmentation map and a geometry-aware level set representation of the target. The level set representation is converted to an approximated segmentation map through a differentiable task transform layer. Simultaneously, we introduce a dual-task consistency regularization between the level set-derived segmentation maps and directly predicted segmentation maps for both labeled and unlabeled data. Extensive experiments on two public datasets show that our method can largely improve the performance by incorporating the unlabeled data. Meanwhile, our framework outperforms the state-of-the-art semi-supervised learning methods.

Authors

Jieneng Chen

5 papers

Guotai Wang

6 papers

Shaoting Zhang

7 papers

TL;DR

Abstract

Authors

References51 items

CPM-Net: A 3D Center-Points Matching Network for Pulmonary Nodule Detection in CT Scans

Learning Geodesic Active Contours for Embedding Object Global Information in Segmentation CNNs

Shape-aware Semi-supervised 3D Semantic Segmentation for Medical Images

Rubik's Cube+: A self-supervised feature learning framework for 3D medical image analysis

Robust Learning Through Cross-Task Consistency

A Global Benchmark of Algorithms for Segmenting Late Gadolinium-Enhanced Cardiac Magnetic Resonance Imaging

Joint Semantic Segmentation and Boundary Detection Using Iterative Pyramid Contexts

Semi-Supervised Semantic Segmentation With Cross-Consistency Training

How Distance Transform Maps Boost Segmentation CNNs: An Empirical Study

Shape-Aware Organ Segmentation by Predicting Signed Distance Maps

Deep Distance Transform for Tubular Structure Segmentation in CT Scans

PyTorch: An Imperative Style, High-Performance Deep Learning Library

HigherHRNet: Scale-Aware Representation Learning for Bottom-Up Human Pose Estimation

Uncertainty-aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentation

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

Transformation-Consistent Self-Ensembling Model for Semisupervised Medical Image Segmentation

Automatic Choroid Layer Segmentation from Optical Coherence Tomography Images Using Deep Learning

FickleNet: Weakly and Semi-Supervised Semantic Image Segmentation Using Stochastic Inference

ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation

3D Semi-Supervised Learning with Uncertainty-Aware Multi-View Co-Training

ASDNet: Attention Based Semi-supervised Deep Networks for Medical Image Segmentation

Taskonomy: Disentangling Task Transfer Learning

Semi-Supervised 3D Abdominal Multi-Organ Segmentation Via Deep Multi-Planar Co-Training

Deep Co-Training for Semi-Supervised Image Recognition

Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation

Interactive Medical Image Segmentation Using Deep Learning With Image-Specific Fine Tuning

Deep Adversarial Networks for Biomedical Image Segmentation Utilizing Unannotated Images

Semi-supervised Learning for Network-Based Cardiac MR Image Segmentation

DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentation

Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results

V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation

Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning

ScribbleSup: Scribble-Supervised Convolutional Networks for Semantic Segmentation

DeepOrgan: Multi-level Deep Convolutional Networks for Automated Pancreas Segmentation

U-Net: Convolutional Networks for Biomedical Image Segmentation

BoxSup: Exploiting Bounding Boxes to Supervise Convolutional Networks for Semantic Segmentation

Fully convolutional networks for semantic segmentation

Two-Stream Convolutional Networks for Action Recognition in Videos

Semi-supervised clustering for MR brain image segmentation

Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials

Segmentation of retinal blood vessels using the radial projection and semi-supervised approach

Level set evolution without re-initialization: a new variational formulation

Consistency-based Semi-supervised Learning for Object detection

A Survey on Medical Image Segmentation

Compared with existing methods

the semi-supervised setting

Sample batch b l = ( x i , y i ) ∈ D l and b = b l + b u , where b u = x i ∈ D u

Output: Dual-task model’s parameter θ 1 for segmentation head, θ 2 for level-set function (LSF) head and θ for shared-weights backbone network

while stopping criterion not met: do

In the fully supervised setting, our dual-task consistency regularization outperforms the separate and joint supervision of dual tasks

f 2 ( x ) = LSF task branch with shared parameter θ and LSF head’s parameter θ 2

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names