Active Learning Classification from a Signal Separation Perspective

TL;DR

This paper introduces a novel signal separation-inspired framework for active learning classification that efficiently identifies class supports in overlapping distributions, demonstrating competitive results on hyperspectral datasets.

Contribution

The paper presents a new clustering and classification framework based on signal separation principles, addressing class support identification in overlapping distributions.

Findings

Competitive performance with state-of-the-art active learning methods

Effective on real-world hyperspectral datasets

Uses fewer training points for high accuracy

Abstract

In machine learning, classification is usually seen as a function approximation problem, where the goal is to learn a function that maps input features to class labels. In this paper, we propose a novel clustering and classification framework inspired by the principles of signal separation. This approach enables efficient identification of class supports, even in the presence of overlapping distributions. We validate our method on real-world hyperspectral datasets Salinas and Indian Pines. The experimental results demonstrate that our method is competitive with the state of the art active learning algorithms by using a very small subset of data set as training points.