Training Methods of Multi-label Prediction Classifiers for Hyperspectral Remote Sensing Images

TL;DR

This paper introduces a multi-label, patch-level deep learning classification method for hyperspectral remote sensing images, comparing three training schemes and highlighting the effectiveness of the joint scheme despite its computational cost.

Contribution

It proposes a novel multi-label classification approach with three training schemes and evaluates their performance on hyperspectral images, emphasizing the joint scheme's effectiveness.

Findings

The joint training scheme yields the best performance.

The iterative scheme is better for complex multi-label data.

Different architectures perform well with the proposed sampling method.

Abstract

With their combined spectral depth and geometric resolution, hyperspectral remote sensing images embed a wealth of complex, non-linear information that challenges traditional computer vision techniques. Yet, deep learning methods known for their representation learning capabilities prove more suitable for handling such complexities. Unlike applications that focus on single-label, pixel-level classification methods for hyperspectral remote sensing images, we propose a multi-label, patch-level classification method based on a two-component deep-learning network. We use patches of reduced spatial dimension and a complete spectral depth extracted from the remote sensing images. Additionally, we investigate three training schemes for our network: Iterative, Joint, and Cascade. Experiments suggest that the Joint scheme is the best-performing scheme; however, its application requires an expensive search for the best weight combination of the loss constituents. The Iterative scheme enables the sharing of features between the two parts of the network at the early stages of training. It performs better on complex data with multi-labels. Further experiments showed that methods designed with different architectures performed well when trained on patches extracted and labeled according to our sampling method.