Chaos-Enhanced Prototypical Networks for Few-Shot Medical Image Classification

2026-04-19

AI Summary

This work addresses the instability of prototypes in few-shot medical image classification, which arises from morphological noise and high intra-class variance. To mitigate this issue, the authors propose a lightweight regularization method based on the logistic chaotic map. For the first time, a deterministic chaotic system is integrated into Prototypical Networks, where controlled chaotic perturbations are injected into support-set features within a fine-tuned ResNet-18 backbone. Leveraging the ergodicity of chaotic dynamics, the method enhances the robustness of the embedding space against noise, thereby stabilizing high-dimensional prototype representations. Evaluated on a 4-way 5-shot brain tumor classification task, the model achieves a test accuracy of 84.52%, significantly outperforming the standard ProtoNet and demonstrating both the effectiveness and novelty of the proposed approach.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Adversarial Attacks & Robustness
• Natural Language Processing: Safety and Robustness

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

The scarcity of labeled clinical data in oncology makes Few-Shot Learning (FSL) a critical framework for Computer Aided Diagnostics, but we observed that standard Prototypical Networks often struggle with the "prototype instability" caused by morphological noise and high intra-class variance in brain tumor scans. Our work attempts to minimize this by integrating a non-linear Logistic Chaos Module into a fine-tuned ResNet-18 backbone creating the Chaos-Enhanced ProtoNet(CE-ProtoNet). Using the deterministic ergodicity of the logistic chaos map we inject controlled perturbations into support features during episodic training-essentially for "stress testing" the embedding space. This process makes the model to converge on noise-invariant representations without increasing computational overhead. Testing this on a 4-way 5-shot brain tumor classification task, we found that a 15% chaotic injection level worked efficiently to stabilize high-dimensional clusters and reduce class dispersion. Our method achieved a peak test accuracy of 84.52%, outperforming standard ProtoNet. Our results suggest the idea of using chaotic perturbation as an efficient, low-overhead regularization tool, for the data-scarce regimes.

Problem

• Few-Shot Learning
• Prototype Instability
• Medical Image Classification
• Brain Tumor
• Intra-class Variance

Innovation

• chaotic perturbation
• prototypical networks
• few-shot learning
• medical image classification
• noise-invariant representation