Color Image Classification via Quaternion Principal Component Analysis Network

TL;DR

This paper introduces QPCANet, an extension of PCANet that incorporates quaternion-based processing to better handle color images, resulting in improved classification accuracy across multiple datasets.

Contribution

The paper proposes QPCANet, a novel quaternion-based deep learning architecture that enhances color image classification by capturing spatial and color information more effectively.

Findings

QPCANet outperforms PCANet on various color image datasets.

QPCANet achieves higher classification accuracy.

Increased intra-class invariance for color images.

Abstract

The Principal Component Analysis Network (PCANet), which is one of the recently proposed deep learning architectures, achieves the state-of-the-art classification accuracy in various databases. However, the performance of PCANet may be degraded when dealing with color images. In this paper, a Quaternion Principal Component Analysis Network (QPCANet), which is an extension of PCANet, is proposed for color images classification. Compared to PCANet, the proposed QPCANet takes into account the spatial distribution information of color images and ensures larger amount of intra-class invariance of color images. Experiments conducted on different color image datasets such as Caltech-101, UC Merced Land Use, Georgia Tech face and CURet have revealed that the proposed QPCANet achieves higher classification accuracy than PCANet.