Partial-Attribution Instance Segmentation for Astronomical Source Detection and Deblending

TL;DR

This paper introduces Partial-Attribution Instance Segmentation, a novel deep learning approach for separating overlapping astronomical sources in complex imaging data, addressing challenges like high dynamic range and low signal-to-noise ratio.

Contribution

The paper presents a new neural network method for source detection and deblending in astronomical images, improving tractability for deep learning models.

Findings

Demonstrated the effectiveness of the proposed neural network approach.

Addressed challenges of high dynamic range and low SNR in astronomical images.

Enabled more accurate source separation in overlapping astronomical objects.

Abstract

Astronomical source deblending is the process of separating the contribution of individual stars or galaxies (sources) to an image comprised of multiple, possibly overlapping sources. Astronomical sources display a wide range of sizes and brightnesses and may show substantial overlap in images. Astronomical imaging data can further challenge off-the-shelf computer vision algorithms owing to its high dynamic range, low signal-to-noise ratio, and unconventional image format. These challenges make source deblending an open area of astronomical research, and in this work, we introduce a new approach called Partial-Attribution Instance Segmentation that enables source detection and deblending in a manner tractable for deep learning models. We provide a novel neural network implementation as a demonstration of the method.