Self-supervised Multi-scale Consistency for Weakly Supervised Segmentation Learning

  title={Self-supervised Multi-scale Consistency for Weakly Supervised Segmentation Learning},
  author={Gabriele Valvano and Andrea Leo and Sotirios A. Tsaftaris},
Collecting large-scale medical datasets with fine-grained annotations is time-consuming and requires experts. For this reason, weakly supervised learning aims at optimising machine learning models using weaker forms of annotations, such as scribbles, which are easier and faster to collect. Unfortunately, training with weak labels is challenging and needs regularisation. Herein, we introduce a novel self-supervised multiscale consistency loss, which, coupled with an attention mechanism… Expand

Figures from this paper

Transferable Time-Series Forecasting under Causal Conditional Shift
This paper analyzes the variational conditional dependencies in time-series data and considers that the causal structures are stable among different domains, and raises the causal conditional shift assumption, and devise an end-to-end model for transferable time- series forecasting. Expand


PGL: Prior-Guided Local Self-supervised Learning for 3D Medical Image Segmentation
An extensive evaluation on four public computerized tomography datasets that cover 11 kinds of major human organs and two tumors indicates that using pre-trained PGL model to initialize a downstream network leads to a substantial performance improvement over both random initialization and the initialization with global consistency-based models. Expand
Temporal Consistency Objectives Regularize the Learning of Disentangled Representations
This work proposes a temporal transformer architecture that given an image conditioned on phase difference, it predicts a future frame and demonstrates that using this regularization, it achieves competitive results and improve semi-supervised segmentation, especially when very few labelled data are available. Expand
Learning to Segment Medical Images with Scribble-Supervision Alone
This work investigates training strategies to learn the parameters of a pixel-wise segmentation network from scribble annotations alone and finds that the networks trained on scribbles suffer from a remarkably small degradation in Dice of only 2.9% and 4.5% with respect to a network trained on full annotations. Expand
Weakly supervised segmentation with cross-modality equivariant constraints
This work presents a novel learning strategy that leverages self-supervision in a multi-modal image scenario to significantly enhance original CAMs, and outperforms relevant recent literature under the same learning conditions. Expand
Multi-Scale Self-Guided Attention for Medical Image Segmentation
  • Ashish Sinha, J. Dolz
  • Computer Science, Medicine
  • IEEE Journal of Biomedical and Health Informatics
  • 2021
Compared to other state-of-the-art segmentation networks, this model yields better segmentation performance, increasing the accuracy of the predictions while reducing the standard deviation, which demonstrates the efficiency of the approach to generate precise and reliable automatic segmentations of medical images. Expand
Prior-Aware Neural Network for Partially-Supervised Multi-Organ Segmentation
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, and is reformulated as a min-max form and optimized via the stochastic primal-dual gradient algorithm. Expand
Deep Interpretable Classification and Weakly-Supervised Segmentation of Histology Images via Max-Min Uncertainty
Comprehensive experiments and ablation studies over the public GlaS colon cancer data and a Camelyon16 patch-based benchmark for breast cancer show substantial improvements over state-of-the-art WSL methods, and confirm the effect of the new regularizers. Expand
Self-Supervised Learning for Cardiac MR Image Segmentation by Anatomical Position Prediction
This paper proposes a novel way for training a cardiac MR image segmentation network, in which features are learnt in a self-supervised manner by predicting anatomical positions, and demonstrates that this seemingly simple task provides a strong signal for feature learning. Expand
Shape-aware Meta-learning for Generalizing Prostate MRI Segmentation to Unseen Domains
A novel shape-aware meta-learning scheme is presented to improve the model generalization in prostate MRI segmentation, and outperforms many state-of-the-art generalization methods consistently across all six settings of unseen domains. Expand
Constrained‐CNN losses for weakly supervised segmentation☆
A differentiable penalty is proposed, which enforces inequality constraints directly in the loss function, avoiding expensive Lagrangian dual iterates and proposal generation and has the potential to close the gap between weakly and fully supervised learning in semantic medical image segmentation. Expand