Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Lifelong Histopathology Whole Slide Image Retrieval via Distance Consistency Rehearsal

Xinyu Zhu, Zhiguo Jiang, Kun Wu, Jun Shi, Yushan Zheng

TL;DR

The paper tackles the problem of catastrophic forgetting in content-based histopathology whole-slide image retrieval (CBHIR) as WSIs databases grow. It introduces Lifelong Whole Slide Retrieval (LWSR), a framework that preserves plasticity with a reservoir-sampled local memory bank and enforces stability for old tasks via a Distance Consistency Rehearsal (DCR) module, optimizing a joint objective that includes pair-wise, cross-entropy, and distance-consistency terms. Empirical results on four TCGA datasets show LWSR achieves retrieval performance close to a joint-training upper bound while outperforming various continual-learning baselines, with improved queue stability as measured by rank-correlations. The approach enables scalable, reliable CBHIR in clinical settings by maintaining consistent retrieval behavior across continual updates, which is crucial for expert-assisted diagnosis.

Abstract

Content-based histopathological image retrieval (CBHIR) has gained attention in recent years, offering the capability to return histopathology images that are content-wise similar to the query one from an established database. However, in clinical practice, the continuously expanding size of WSI databases limits the practical application of the current CBHIR methods. In this paper, we propose a Lifelong Whole Slide Retrieval (LWSR) framework to address the challenges of catastrophic forgetting by progressive model updating on continuously growing retrieval database. Our framework aims to achieve the balance between stability and plasticity during continuous learning. To preserve system plasticity, we utilize local memory bank with reservoir sampling method to save instances, which can comprehensively encompass the feature spaces of both old and new tasks. Furthermore, A distance consistency rehearsal (DCR) module is designed to ensure the retrieval queue's consistency for previous tasks, which is regarded as stability within a lifelong CBHIR system. We evaluated the proposed method on four public WSI datasets from TCGA projects. The experimental results have demonstrated the proposed method is effective and is superior to the state-of-the-art methods.

Lifelong Histopathology Whole Slide Image Retrieval via Distance Consistency Rehearsal

TL;DR

The paper tackles the problem of catastrophic forgetting in content-based histopathology whole-slide image retrieval (CBHIR) as WSIs databases grow. It introduces Lifelong Whole Slide Retrieval (LWSR), a framework that preserves plasticity with a reservoir-sampled local memory bank and enforces stability for old tasks via a Distance Consistency Rehearsal (DCR) module, optimizing a joint objective that includes pair-wise, cross-entropy, and distance-consistency terms. Empirical results on four TCGA datasets show LWSR achieves retrieval performance close to a joint-training upper bound while outperforming various continual-learning baselines, with improved queue stability as measured by rank-correlations. The approach enables scalable, reliable CBHIR in clinical settings by maintaining consistent retrieval behavior across continual updates, which is crucial for expert-assisted diagnosis.

Abstract

Content-based histopathological image retrieval (CBHIR) has gained attention in recent years, offering the capability to return histopathology images that are content-wise similar to the query one from an established database. However, in clinical practice, the continuously expanding size of WSI databases limits the practical application of the current CBHIR methods. In this paper, we propose a Lifelong Whole Slide Retrieval (LWSR) framework to address the challenges of catastrophic forgetting by progressive model updating on continuously growing retrieval database. Our framework aims to achieve the balance between stability and plasticity during continuous learning. To preserve system plasticity, we utilize local memory bank with reservoir sampling method to save instances, which can comprehensively encompass the feature spaces of both old and new tasks. Furthermore, A distance consistency rehearsal (DCR) module is designed to ensure the retrieval queue's consistency for previous tasks, which is regarded as stability within a lifelong CBHIR system. We evaluated the proposed method on four public WSI datasets from TCGA projects. The experimental results have demonstrated the proposed method is effective and is superior to the state-of-the-art methods.
Paper Structure (13 sections, 3 equations, 2 figures, 5 tables, 1 algorithm)

This paper contains 13 sections, 3 equations, 2 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Method
  3. Data Preparation and Problem Definition
  4. Distance Consistency Rehearsal
  5. Objective for LWSR
  6. Experiment and Result
  7. Model verification
  8. Comparison with other continuous learning frameworks
  9. Conclusion
  10. Acknowledgments.
  11. Disclosure of Interests.
  12. Evaluation metrics
  13. Implementation details

Figures (2)

  • Figure 1: The overview of the proposed lifelong whole slide retrieval (LWSR) framework, where (I) shows the ever-growing histopathology image database, (II) is the reservoir random sampling method for buffer sampling, (III) describe universal training process of every task, and (IV) illutrates the proposed distance consistency rehearsal (DCR) module that is detailed in section \ref{['section_2_2']} and algorithm \ref{['algorithm_dcr']}.
  • Figure 2: Visualization of the capacity of the proposed method in keeping returned queues consistent after continual learning, inconsistent returned WSIs are framed in red and irrelevant results are filled with blue.