Approximate nonnegative matrix factorization algorithm for the analysis of angular differential imaging data

TL;DR

This paper introduces an approximate nonnegative matrix factorization algorithm tailored for analyzing angular differential imaging data, enhancing the detection of faint astronomical objects like exoplanets.

Contribution

It proposes a novel approximate NMF algorithm specifically designed for ADI data, improving contrast and detection capabilities in high-resolution astronomical imaging.

Findings

Enhanced contrast in ADI data using the proposed NMF algorithm

Improved detection of faint exoplanets in high-resolution images

Demonstrated effectiveness over PCA-based methods

Abstract

The angular differential imaging (ADI) is used to improve contrast in high resolution astronomical imaging. An example is the direct imaging of exoplanet in camera fed by Extreme Adaptive Optics. The subtraction of the main dazzling object to observe the faint companion was improved using Principal Component Analysis (PCA). It factorizes the positive astronomical frames into positive and negative components. On the contrary, the Nonnegative Matrix Factorization (NMF) uses only positive components, mimicking the actual composition of the long exposure images.