LGNov 10, 2020

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

arXiv:2011.05384v14 citations
AI Analysis

This work addresses the need for efficient factor analysis in streaming data scenarios for applications like image and time-series processing, but it appears incremental as it applies existing ONMF methods to new contexts.

The paper tackled the problem of learning joint dictionary atoms from correlated datasets using online nonnegative matrix factorization (ONMF), proposing a temporal dictionary learning scheme for time-series data and demonstrating it on historical temperature data, video frames, and color images.

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.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes