Applications of Online Nonnegative Matrix Factorization to Image and Time-Series Data

TL;DR

This paper explores the use of online nonnegative matrix factorization (ONMF) for real-time analysis of image and time-series data, enabling concurrent learning from streaming data in various applications.

Contribution

The paper introduces a temporal dictionary learning scheme using ONMF for time-series data and demonstrates its effectiveness on temperature, video, and image datasets.

Findings

Effective joint dictionary learning from correlated datasets

Successful application to temperature, video, and image data

Real-time data processing capabilities

Abstract

Online nonnegative matrix factorization (ONMF) is a matrix factorization technique in the online setting where data are acquired in a streaming fashion and the matrix factors are updated each time. This enables factor analysis to be performed concurrently with the arrival of new data samples. In this article, we demonstrate how one can use online nonnegative matrix factorization algorithms to learn joint dictionary atoms from an ensemble of correlated data sets. We propose a temporal dictionary learning scheme for time-series data sets, based on ONMF algorithms. We demonstrate our dictionary learning technique in the application contexts of historical temperature data, video frames, and color images.