Deep neural networks with Fisher vector encoding for medical image classification

TL;DR

This paper introduces a Fisher Vector encoding method integrated with hybrid CNN + ViT models for medical image classification, improving performance across datasets of varying sizes.

Contribution

It proposes a novel approach combining Fisher Vectors with hybrid CNN + ViT architectures, addressing dataset size scalability and computational efficiency.

Findings

Outperforms benchmarks on all MedMNIST (v2) datasets.

Achieves competitive results on Clean-CC-CCII and ISIC2018.

Proposes a cost-limiting GMM estimation method for large datasets.

Abstract

Orderless encoding methods have shown to improve Convolutional Neural Networks (CNNs) for image classification in the context of limited availability of data. Additionally, hybrid CNN + Vision Transformers (ViT) models have been recently proposed to address CNN locality bias issues. These models outperformed CNN-only approaches. Despite that, the integration of such hybrid models with more elaborated feature representation can be highly beneficial and remains large unexplored in the literature. In this context, we propose the introduction of an orderless encoding method, Fisher Vectors, to hybrid CNN + ViT architectures, aiming at achieving a model suitable for both small and large datasets. Such enconding method relies on estimating a Gaussian Mixture Model (GMM) on image features. In large datasets, computational costs of the GMM estimation is a limiting factor for the application of Fisher Vectors. Thus, we propose a method to limit the growth of GMM estimation costs as we increase the size of the dataset. We explore the feasibility of our method in the context of medical image classification by appling it to MedMNIST (v2), Clean-CC-CCII and ISIC2018. This collection of datasets contains a wide variety of data scales and modalities. We outperform benchmark results in all MedMNIST (v2) datasets and obtain literature-competitive results in Clean-CC-CCII and ISIC2018.