Salient Slices: Improved Neural Network Training and Performance with Image Entropy

TL;DR

This paper introduces a method that slices images into entropy-based tiles for CNN training and prediction, improving accuracy, enabling high-resolution analysis, and aiding data augmentation especially with limited data.

Contribution

It proposes using image entropy to select informative image slices, enhancing CNN training and prediction for high-resolution images.

Findings

Improved CNN accuracy with entropy-based image slicing

Effective analysis of high-resolution images

Enhanced data augmentation for limited datasets

Abstract

As a training and analysis strategy for convolutional neural networks (CNNs), we slice images into tiled segments and use, for training and prediction, segments that both satisfy a criterion of information diversity and contain sufficient content to support classification. In particular, we utilize image entropy as the diversity criterion. This ensures that each tile carries as much information diversity as the original image, and for many applications serves as an indicator of usefulness in classification. To make predictions, a probability aggregation framework is applied to probabilities assigned by the CNN to the input image tiles. This technique facilitates the use of large, high-resolution images that would be impractical to analyze unmodified; provides data augmentation for training, which is particularly valuable when image availability is limited; and the ensemble nature of the input for prediction enhances its accuracy.