Divergence Regulated Encoder Network for Joint Dimensionality Reduction and Classification

TL;DR

This paper introduces a divergence regulated encoder network that jointly reduces dimensionality and classifies remote-sensing images, improving interpretability and transferability of features while maintaining high accuracy.

Contribution

It proposes a novel histogram neural network inspired by t-SNE for joint dimensionality reduction and classification, with divergence measures enhancing embedding quality.

Findings

Maintains or improves classification accuracy

Embeds out-of-sample points effectively

Reduces feature dimensionality while preserving class discriminability

Abstract

Feature representation is an important aspect of remote-sensing based image classification. While deep convolutional neural networks are able to effectively amalgamate information, large numbers of parameters often make learned features inscrutable and difficult to transfer to alternative models. In order to better represent statistical texture information for remote-sensing image classification, in this paper, we investigate performing joint dimensionality reduction and classification using a novel histogram neural network. Motivated by a popular dimensionality reduction approach, t-Distributed Stochastic Neighbor Embedding (t-SNE), our proposed method incorporates a classification loss computed on samples in a low-dimensional embedding space. We compare the learned sample embeddings against coordinates found by t-SNE in terms of classification accuracy and qualitative assessment. We also explore use of various divergence measures in the t-SNE objective. The proposed method has several advantages such as readily embedding out-of-sample points and reducing feature dimensionality while retaining class discriminability. Our results show that the proposed approach maintains and/or improves classification performance and reveals characteristics of features produced by neural networks that may be helpful for other applications.