Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cross-Domain Evaluation of Few-Shot Classification Models: Natural Images vs. Histopathological Images

Ardhendu Sekhar, Aditya Bhattacharya, Vinayak Goyal, Vrinda Goel, Aditya Bhangale, Ravi Kant Gupta, Amit Sethi

TL;DR

The performance of few-shot classification models across different domains, specifically natural images and histopathological images are investigated and insights into the transferability and generalization capabilities of few-shot classification models between diverse image domains are revealed.

Abstract

In this study, we investigate the performance of few-shot classification models across different domains, specifically natural images and histopathological images. We first train several few-shot classification models on natural images and evaluate their performance on histopathological images. Subsequently, we train the same models on histopathological images and compare their performance. We incorporated four histopathology datasets and one natural images dataset and assessed performance across 5-way 1-shot, 5-way 5-shot, and 5-way 10-shot scenarios using a selection of state-of-the-art classification techniques. Our experimental results reveal insights into the transferability and generalization capabilities of few-shot classification models between diverse image domains. We analyze the strengths and limitations of these models in adapting to new domains and provide recommendations for optimizing their performance in cross-domain scenarios. This research contributes to advancing our understanding of few-shot learning in the context of image classification across diverse domains.

Cross-Domain Evaluation of Few-Shot Classification Models: Natural Images vs. Histopathological Images

TL;DR

The performance of few-shot classification models across different domains, specifically natural images and histopathological images are investigated and insights into the transferability and generalization capabilities of few-shot classification models between diverse image domains are revealed.

Abstract

In this study, we investigate the performance of few-shot classification models across different domains, specifically natural images and histopathological images. We first train several few-shot classification models on natural images and evaluate their performance on histopathological images. Subsequently, we train the same models on histopathological images and compare their performance. We incorporated four histopathology datasets and one natural images dataset and assessed performance across 5-way 1-shot, 5-way 5-shot, and 5-way 10-shot scenarios using a selection of state-of-the-art classification techniques. Our experimental results reveal insights into the transferability and generalization capabilities of few-shot classification models between diverse image domains. We analyze the strengths and limitations of these models in adapting to new domains and provide recommendations for optimizing their performance in cross-domain scenarios. This research contributes to advancing our understanding of few-shot learning in the context of image classification across diverse domains.

Paper Structure

This paper contains 12 sections, 27 equations, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Prototypical Networks
  5. SimpleShot
  6. Laplacianshot
  7. DeepEMD
  8. DeepBDC
  9. Experiments and Results
  10. Dataset
  11. Results and Discussion
  12. Conclusion

Figures (5)

  • Figure 1: The diagram depicts a few-shot learning model, demonstrating its capability to generalize effectively and recognize classes within an unlabeled query set using only a limited number of support examples. In the illustration, five different colors in the support set represent five distinct classes (ways), each having one sample (shot). fewshotbioimaging
  • Figure 2: The NCT dataset is illustrated by the sample images in the top two rows, whereas the CRC-TP dataset is represented by the sample images in the bottom row. fewshotbioimaging
  • Figure 3: 5-way 1-shot 2-query episode with support set images from five LC25000 dataset classes in the first row and query set images in the last two rows. fewshotbioimaging
  • Figure 4: Performance of DeepEMD on crc-tp dataset with different training conditions
  • Figure 5: Performance of laplacianshot on nct dataset with different training conditions