Semi-Supervised Deep Learning for Fully Convolutional Networks

Abstract

Deep learning usually requires large amounts of labeled training data, but annotating data is costly and tedious. The framework of semi-supervised learning provides the means to use both labeled data and arbitrary amounts of unlabeled data for training. Recently, semi-supervised deep learning has been intensively studied for standard CNN architectures. However, Fully Convolutional Networks (FCNs) set the state-of-the-art for many image segmentation tasks. To the best of our knowledge, there is no existing semi-supervised learning method for such FCNs yet. We lift the concept of auxiliary manifold embedding for semi-supervised learning to FCNs with the help of Random Feature Embedding. In our experiments on the challenging task of MS Lesion Segmentation, we leverage the proposed framework for the purpose of domain adaptation and report substantial improvements over the baseline model.

9 pages, 6 figures

Cited by in corpus (23)

• Medical Image Segmentation Using Deep Learning: A Survey
• Shape-aware Semi-supervised 3D Semantic Segmentation for Medical Images
• Embracing Imperfect Datasets: A Review of Deep Learning Solutions for Medical Image Segmentation
• Semi-supervised Medical Image Classification with Relation-driven Self-ensembling Model
• Semi-Supervised Medical Image Segmentation via Learning Consistency under Transformations
• Uncertainty-aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentation
• Curriculum semi-supervised segmentation
• Factorised spatial representation learning: application in semi-supervised myocardial segmentation
• Domain and Geometry Agnostic CNNs for Left Atrium Segmentation in 3D Ultrasound
• Deep semi-supervised segmentation with weight-averaged consistency targets
• BoxNet: Deep Learning Based Biomedical Image Segmentation Using Boxes Only Annotation
• Revisiting CycleGAN for semi-supervised segmentation
• Boosting Semi-supervised Image Segmentation with Global and Local Mutual Information Regularization
• Deep Co-Training for Semi-Supervised Image Segmentation
• Semi-Supervised Brain Lesion Segmentation with an Adapted Mean Teacher Model
• Extreme Consistency: Overcoming Annotation Scarcity and Domain Shifts
• ROAM: Random Layer Mixup for Semi-Supervised Learning in Medical Imaging
• Deep Semi-supervised Knowledge Distillation for Overlapping Cervical Cell Instance Segmentation
• Anatomical Data Augmentation via Fluid-based Image Registration
• Multiple Sclerosis Lesion Segmentation -- A Survey of Supervised CNN-Based Methods
• A generic ensemble based deep convolutional neural network for semi-supervised medical image segmentation
• Improving Calibration and Out-of-Distribution Detection in Medical Image Segmentation with Convolutional Neural Networks
• Uncertainty guided semi-supervised segmentation of retinal layers in OCT images