Compound Figure Separation of Biomedical Images: Mining Large Datasets for Self-supervised Learning

  title={Compound Figure Separation of Biomedical Images: Mining Large Datasets for Self-supervised Learning},
  author={Tianyuan Yao and Changbing Qu and Jun Long and Quan Liu and Ruining Deng and Yuan Tian and Jiachen Xu and Aadarsh Jha and Zuhayr Asad and Shunxing Bao and Mengyang Zhao and Agnes B. Fogo and Bennett A.Landman and Haichun Yang and Catie Chang and Yuankai Huo},
With the rapid development of self-supervised learning (e.g., contrastive learning), the im-portance of having large-scale images (even without annotations) for training a more generalizable AI model has been widely recognized in medical image analysis. However, collecting large-scale task-specific unannotated data at scale can be challenging for individual labs. Existing online resources, such as digital books, publications, and search engines, provide a new resource for obtaining large-scale… 

Figures and Tables from this paper



Exploring Simple Siamese Representation Learning

  • Xinlei ChenKaiming He
  • Computer Science
    2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • 2021
Surprising empirical results are reported that simple Siamese networks can learn meaningful representations even using none of the following: (i) negative sample pairs, (ii) large batches, (iii) momentum encoders.

A Simple Framework for Contrastive Learning of Visual Representations

It is shown that composition of data augmentations plays a critical role in defining effective predictive tasks, and introducing a learnable nonlinear transformation between the representation and the contrastive loss substantially improves the quality of the learned representations, and contrastive learning benefits from larger batch sizes and more training steps compared to supervised learning.

Compound Figure Separation of Biomedical Images with Side Loss

A simple compound figure separation (SimCFS) framework that uses weak classification annotations from individual images, and achieves a new state-of-the-art performance on the ImageCLEF 2016 Compound Figure Separation Database.

Weighted boxes fusion: Ensembling boxes from different object detection models

AI Applications in Renal Pathology.

A Two-Stage Framework for Compound Figure Separation

A new strategy for compound Figure separation is proposed, which decomposes the compound figures into constituent subfigures while preserving the association between the subfigURES and their respective caption components.

Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning

This work introduces Bootstrap Your Own Latent (BYOL), a new approach to self-supervised image representation learning that performs on par or better than the current state of the art on both transfer and semi- supervised benchmarks.

Chart Mining: A Survey of Methods for Automated Chart Analysis

A comprehensive survey of approaches across all components of the automated chart mining pipeline such as automated extraction of charts from documents, processing of multi-panel charts, and datasets for training and evaluation are presented.

YOLOv4: Optimal Speed and Accuracy of Object Detection

This work uses new features: WRC, CSP, CmBN, SAT, Mish activation, Mosaic data augmentation, C mBN, DropBlock regularization, and CIoU loss, and combine some of them to achieve state-of-the-art results: 43.5% AP for the MS COCO dataset at a realtime speed of ~65 FPS on Tesla V100.

Unified deep neural network for segmentation and labeling of multipanel biomedical figures

A deep convolutional neural network is proposed, which splits the panels and recognizes the panel labels in a single step and is evaluated on the ImageCLEF data set and achieved better performance than the results reported in the literature.