# Dynamic and polarimetric VLBI imaging with a multiscalar approach

**Authors:** Hendrik Müller, Andrei Lobanov  
**arXiv:** [2303.11877](https://arxiv.org/abs/2303.11877)  
**Date:** 2023-03-21  
**Summary page:** https://www.summarizepaper.com/en/paper/2303.11877v1/

## Key points

- The paper presents a new approach to dynamic and polarimetric imaging in Very Long Baseline Interferometry (VLBI) using a multiscalar approach called mr-support imaging.
- VLBI is a technique that uses the correlation of signals recorded at single antennas to achieve high spatial resolution, and mr-support imaging aims to recover images of the shadow of supermassive black holes.
- The Event Horizon Telescope (EHT) has achieved the angular resolution needed to capture the first image of the black hole shadow in M87 and in the Milky Way, and the next-generation Event Horizon Telescope (ngEHT) is designed to produce movies of accretion onto central black holes at event horizon scales.
- The proposed method extends previous multiscale imaging approaches such as DoG-HiT, which showed promising performance for solving VLBI imaging problems.
- The mr-support imaging approach utilizes a multiscalar approach by decomposing time-averaged Stokes I images into single subbands using wavelet transforms.
- The authors demonstrate on synthetic data that mr-support imaging offers ample regularization and can recover simple geometric dynamics at horizon scales typical of an EHT setup. They also benchmarked its performance for denser ngEHT configurations, showing major improvements with additional ngEHT antennas for dynamic polarimetric reconstructions.
- The proposed method adds an additional regularizer to existing state-of-the-art dynamic reconstruction methods: constraining reconstruction to the multiresolution support. This simpler regularizer can capture motion with temporal regularizers and priors while offering ample regularization.
- Overall, this work provides an important step towards polarimetric imaging of dynamically evolving structural patterns with high spatial and temporal resolution using current and upcoming instruments like EHT and ngEHT.
- The proposed method is relatively lightweight, fast, and largely automatic and data-driven, making it a promising approach for future VLBI imaging problems.

## Summary

This paper presents a new approach to dynamic and polarimetric imaging in Very Long Baseline Interferometry (VLBI) using a multiscalar approach called mr-support imaging. VLBI is a technique that uses the correlation of signals recorded at single antennas to achieve high spatial resolution, and mr-support imaging aims to recover images of the shadow of supermassive black holes. The Event Horizon Telescope (EHT) is a globally spanning VLBI array that has achieved the angular resolution needed to capture the first image of the black hole shadow in M87 and in the Milky Way. The next-generation Event Horizon Telescope (ngEHT) is a planned extension of the EHT designed to produce movies of accretion onto central black holes at event horizon scales. The proposed method extends previous multiscale imaging approaches such as DoG-HiT, which showed promising performance for solving VLBI imaging problems. The mr-support imaging approach utilizes a multiscalar approach by decomposing time-averaged Stokes I images into single subbands using wavelet transforms. The set of statistically significant wavelet coefficients, or multiresolution support, computed by DoG-HiT serves as a prior for constrained minimization during reconstruction. The authors demonstrate on synthetic data that mr-support imaging offers ample regularization and can recover simple geometric dynamics at horizon scales typical of an EHT setup. They also benchmarked its performance for denser ngEHT configurations, showing major improvements with additional ngEHT antennas for dynamic polarimetric reconstructions. The proposed method adds an additional regularizer to existing state-of-the-art dynamic reconstruction methods: constraining reconstruction to the multiresolution support. This simpler regularizer can capture motion with temporal regularizers and priors while offering ample regularization. Overall, this work provides an important step towards polarimetric imaging of dynamically evolving structural patterns with high spatial and temporal resolution using current and upcoming instruments like EHT and ngEHT. The proposed method is relatively lightweight, fast, and largely automatic and data-driven, making it a promising approach for future VLBI imaging problems.