A Location-Sensitive Local Prototype Network For Few-Shot Medical Image Segmentation

@article{Yu2021ALL,
  title={A Location-Sensitive Local Prototype Network For Few-Shot Medical Image Segmentation},
  author={Qinji Yu and K. Dang and Nima Tajbakhsh and Demetri Terzopoulos and Xiaowei Ding},
  journal={2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)},
  year={2021},
  pages={262-266}
}
  • Qinji Yu, K. Dang, Xiaowei Ding
  • Published 18 March 2021
  • Computer Science
  • 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)
Despite the tremendous success of deep neural networks in medical image segmentation, they typically require a large amount of costly, expert-level annotated data. Few-shot segmentation approaches address this issue by learning to transfer knowledge from limited quantities of labeled examples. Incorporating appropriate prior knowledge is critical in designing high-performance few-shot segmentation algorithms. Since strong spatial priors exist in many medical imaging modalities, we propose a… 

Figures and Tables from this paper

Prototypical few-shot segmentation for cross-institution male pelvic structures with spatial registration

The prowess that makes few-shot learning desirable in medical image analysis is the e ffi cient use of the support image data, which are labelled to classify or segment new classes, a task that

FEW-SHOT Image Segmentation for Cross-Institution Male Pelvic Organs Using Registration-Assisted Prototypical Learning

TLDR
This work presents the first 3D few-shot interclass segmentation network for medical images, using a labelled multi-institution dataset from prostate cancer patients with eight regions of interest, and proposes an image alignment module registering the predicted segmentation of both query and support data to a reference atlas space.

Unsupervised Video Object Segmentation via Prototype Memory Network

TLDR
A novel prototype memory network architecture that effectively extracts the RGB and motion information by extracting superpixel-based component prototypes from the input RGB images and optical maps and adaptively stores the most useful prototypes in memory and discards obsolete prototypes.

Coarse Retinal Lesion Annotations Refinement via Prototypical Learning

TLDR
An annotation refinement network to convert a coarse annotation into a pixel-level segmentation mask and a prototype weighing module to handle chal-lenging cases where the lesion is overly small is proposed.

Self-Supervised Learning for Few-Shot Medical Image Segmentation

TLDR
This work proposes a novel self-supervised FSS framework for medical images, named SSL-ALPNet, in order to bypass the requirement for annotations during training, and exploits superpixel-based pseudo-labels to provide supervision signals.

Bilateral-ViT For Robust Fovea Localization

TLDR
This paper proposes a novel Vision Transformer (ViT) approach that integrates information both inside and outside the fovea region to achieve robust foveo localization and establishes the new state of the arts using the Messidor and PALM datasets.

PoissonSeg: Semi-Supervised Few-Shot Medical Image Segmentation via Poisson Learning

TLDR
A novel semi-supervised FSS framework for medical image segmentation that employs Poisson learning for modeling data relationship and propagating supervision signals, and Spatial Consistency Calibration for encouraging the model to learn more coherent representations is proposed.

References

SHOWING 1-10 OF 20 REFERENCES

Self-Supervision with Superpixels: Training Few-shot Medical Image Segmentation without Annotation

TLDR
A novel self-supervised FSS framework for medical images in order to eliminate the requirement for annotations during training, and superpixel-based pseudo-labels are generated to provide supervision.

PANet: Few-Shot Image Semantic Segmentation With Prototype Alignment

TLDR
This paper tackles the challenging few-shot segmentation problem from a metric learning perspective and presents PANet, a novel prototype alignment network to better utilize the information of the support set to better generalize to unseen object categories.

CANet: Classagnostic segmentation networks with iterative refinement and attentive few-shot learning

  • IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 5217–5226.
  • 2019

Learning Semantics-enriched Representation via Self-discovery, Self-classification, and Self-restoration

TLDR
This work trains deep models to learn semantically enriched visual representation by self-discovery, self-classification, and self-restoration of the anatomy underneath medical images, resulting in a semantics-enriched, general-purpose, pre-trained 3D model, named Semantic Genesis.

Prototype Refinement Network for Few-Shot Segmentation

TLDR
A Prototype Refinement Network (PRNet) is proposed which fuses support prototypes with query prototypes, incorporating the knowledge from both sides and is effective in prototype refinement without importing extra learnable parameters.

Data Augmentation Using Learned Transformations for One-Shot Medical Image Segmentation

TLDR
This work learns a model of transformations from the images, and uses the model along with the labeled example to synthesize additional labeled examples, enabling the synthesis of complex effects such as variations in anatomy and image acquisition procedures.

Surrogate Supervision for Medical Image Analysis: Effective Deep Learning From Limited Quantities of Labeled Data

TLDR
This research employs 3 surrogate supervision schemes, namely rotation, reconstruction, and colorization, in 4 different medical imaging applications representing classification and segmentation for both 2D and 3D medical images, indicating the practical value of abundant unlabeled medical image data.

Few-shot 3D Multi-modal Medical Image Segmentation using Generative Adversarial Learning

TLDR
This work proposes a novel method based on Generative Adversarial Networks (GANs) to train a segmentation model with both labeled and unlabeled images, which prevents over-fitting by learning to discriminate between true and fake patches obtained by a generator network.