Efficient Self-Supervision using Patch-based Contrastive Learning for Histopathology Image Segmentation
Research output: Working paper › Preprint › Research
Standard
Efficient Self-Supervision using Patch-based Contrastive Learning for Histopathology Image Segmentation. / Boserup, Nicklas; Selvan, Raghavendra.
arxiv.org, 2022.Research output: Working paper › Preprint › Research
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - UNPB
T1 - Efficient Self-Supervision using Patch-based Contrastive Learning for Histopathology Image Segmentation
AU - Boserup, Nicklas
AU - Selvan, Raghavendra
N1 - 15 pages, 8 figures. Source code at https://github.com/nickeopti/bach-contrastive-segmentation
PY - 2022
Y1 - 2022
N2 - Learning discriminative representations of unlabelled data is a challenging task. Contrastive self-supervised learning provides a framework to learn meaningful representations using learned notions of similarity measures from simple pretext tasks. In this work, we propose a simple and efficient framework for self-supervised image segmentation using contrastive learning on image patches, without using explicit pretext tasks or any further labeled fine-tuning. A fully convolutional neural network (FCNN) is trained in a self-supervised manner to discern features in the input images and obtain confidence maps which capture the network's belief about the objects belonging to the same class. Positive- and negative- patches are sampled based on the average entropy in the confidence maps for contrastive learning. Convergence is assumed when the information separation between the positive patches is small, and the positive-negative pairs is large. We evaluate this method for the task of segmenting nuclei from multiple histopathology datasets, and show comparable performance with relevant self-supervised and supervised methods. The proposed model only consists of a simple FCNN with 10.8k parameters and requires about 5 minutes to converge on the high resolution microscopy datasets, which is orders of magnitude smaller than the relevant self-supervised methods to attain similar performance.
AB - Learning discriminative representations of unlabelled data is a challenging task. Contrastive self-supervised learning provides a framework to learn meaningful representations using learned notions of similarity measures from simple pretext tasks. In this work, we propose a simple and efficient framework for self-supervised image segmentation using contrastive learning on image patches, without using explicit pretext tasks or any further labeled fine-tuning. A fully convolutional neural network (FCNN) is trained in a self-supervised manner to discern features in the input images and obtain confidence maps which capture the network's belief about the objects belonging to the same class. Positive- and negative- patches are sampled based on the average entropy in the confidence maps for contrastive learning. Convergence is assumed when the information separation between the positive patches is small, and the positive-negative pairs is large. We evaluate this method for the task of segmenting nuclei from multiple histopathology datasets, and show comparable performance with relevant self-supervised and supervised methods. The proposed model only consists of a simple FCNN with 10.8k parameters and requires about 5 minutes to converge on the high resolution microscopy datasets, which is orders of magnitude smaller than the relevant self-supervised methods to attain similar performance.
KW - cs.CV
KW - cs.LG
U2 - 10.48550/arXiv.2208.10779
DO - 10.48550/arXiv.2208.10779
M3 - Preprint
BT - Efficient Self-Supervision using Patch-based Contrastive Learning for Histopathology Image Segmentation
PB - arxiv.org
ER -
ID: 319063582