An Analysis on Ensemble Learning optimized Medical Image Classification with Deep Convolutional Neural Networks

TL;DR

This paper evaluates the impact of various ensemble learning strategies on deep convolutional neural network performance in medical image classification, demonstrating significant improvements in accuracy and robustness across multiple datasets.

Contribution

It introduces a reproducible pipeline analyzing the effects of Augmenting, Stacking, and Bagging ensemble techniques with 9 CNN architectures on medical imaging datasets.

Findings

Stacking yields up to 13% F1-score improvement.

Augmenting consistently improves performance by up to 4%.

Bagging increases F1-score by up to 11%.

Abstract

Novel and high-performance medical image classification pipelines are heavily utilizing ensemble learning strategies. The idea of ensemble learning is to assemble diverse models or multiple predictions and, thus, boost prediction performance. However, it is still an open question to what extent as well as which ensemble learning strategies are beneficial in deep learning based medical image classification pipelines. In this work, we proposed a reproducible medical image classification pipeline for analyzing the performance impact of the following ensemble learning techniques: Augmenting, Stacking, and Bagging. The pipeline consists of state-of-the-art preprocessing and image augmentation methods as well as 9 deep convolution neural network architectures. It was applied on four popular medical imaging datasets with varying complexity. Furthermore, 12 pooling functions for combining multiple predictions were analyzed, ranging from simple statistical functions like unweighted averaging up to more complex learning-based functions like support vector machines. Our results revealed that Stacking achieved the largest performance gain of up to 13% F1-score increase. Augmenting showed consistent improvement capabilities by up to 4% and is also applicable to single model based pipelines. Cross-validation based Bagging demonstrated significant performance gain close to Stacking, which resulted in an F1-score increase up to +11%. Furthermore, we demonstrated that simple statistical pooling functions are equal or often even better than more complex pooling functions. We concluded that the integration of ensemble learning techniques is a powerful method for any medical image classification pipeline to improve robustness and boost performance.