🔭 astrophysics

Informational coherence geometry in astronomical imaging

This paper introduces an informational coherence-geometry pipeline that demonstrates how phase-sensitive topological analysis of astronomical images reveals significantly larger connected high-coherence regions in real data compared to phase-randomized surrogates, thereby establishing coherence topology as a distinct and valuable descriptor of spatial organization beyond standard spectral statistics.

Original authors: Raoul Bianchetti, payam Danesh

Published 2026-08-10

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Raoul Bianchetti, payam Danesh

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are looking at a giant, starry painting. For decades, scientists have had a favorite tool to study these cosmic masterpieces: the Fourier transform. Think of this tool as a magical prism that breaks the image down into its pure colors of light and dark, telling us exactly how much energy exists at every size of pattern, from tiny specks to huge swirls. It's like knowing the exact recipe of a cake just by weighing the flour, sugar, and eggs. But here's the catch: knowing the ingredients doesn't tell you how they are arranged. You could have the exact same amount of flour and sugar, but if you mix them up randomly, you get a mess instead of a delicious cake. In the world of space images, this "mixing" is called the phase. The phase is the secret instruction manual that tells the stars and galaxies where to stand to form shapes, lines, and connections. Without it, you have all the energy, but no structure.

This brings us to a fascinating question that has puzzled image scientists for a long time: If we keep the exact same "ingredients" (the energy amounts) but scramble the "instructions" (the phase), does the picture still hold together? In other words, is the way stars and gas connect to each other in real space just a random accident of energy, or is there a deeper, hidden geometry holding them together? This paper dives into that mystery using the incredible new cameras on the James Webb Space Telescope (JWST). It asks a simple but profound question: If we take a real photo of the universe, scramble the hidden instructions, and keep the energy the same, will the scrambled version look just as organized as the real one?

The Great Cosmic Scramble

The authors, Raoul Bianchetti and Payam Danesh, decided to play a game of "spot the difference" with the universe. They didn't just look at the pictures; they built a mathematical pipeline to test if the connections in space are special. Their method is like a chef who takes a perfect, complex dish, measures out every single grain of salt and spice, and then tries to recreate the dish by throwing those exact ingredients into a blender and hoping they magically reassemble into the original meal.

Here is how they did it:

The Real Deal: They took two stunning, public photos from the JWST. One was the COSMOS Legacy Field, a deep patch of sky filled with thousands of faint, distant galaxies and wispy gas. The other was Cassiopeia A, a nearby exploded star (supernova remnant) that looks like a giant, glowing shell.
The Scramble: For each photo, they used a computer trick to create a "surrogate" image. They kept the exact same energy levels (the Fourier amplitude) but completely randomized the phase. Imagine taking a puzzle, keeping all the pieces, but shuffling them so the picture is gone, yet the total color and brightness of the pile remain identical.
The Test: They then looked for "coherence." In simple terms, coherence is how well the image parts stick together in a line or a shape. They asked: "Do the bright, connected, filament-like structures in the real photo form big, long chains, or do they break apart into tiny, scattered islands?"

The Big Discovery: Real vs. Random

The results were surprisingly clear. When they looked at the real images, the "sticky" parts of the universe formed massive, connected highways. In the COSMOS field, the largest connected group of these organized pixels was a giant cluster of 1,358 pixels long. It was a massive, continuous thread of structure.

Now, look at the scrambled versions. Even though they had the exact same amount of light and energy, the structures fell apart. The largest connected group in the scrambled COSMOS images was only 223 pixels. That's a tiny fraction of the real thing! The scrambled images had more little clusters, but they were all short, broken, and messy. The real universe, however, had organized itself into something much bigger and more connected.

They did the same test with Cassiopeia A. The real image showed a massive connected shell of 780 pixels. The scrambled versions? Their biggest cluster was only 285 pixels.

To make sure this wasn't just a lucky fluke, they ran the scramble 50 times for each image. In every single case, the real image's biggest cluster was larger than the biggest cluster in any of the 50 scrambled versions. The odds of this happening by chance are about 1 in 51 (or roughly 0.02). This suggests that the way the universe organizes itself is not just a random arrangement of energy; there is a specific "phase" geometry that holds these structures together.

What This Means (and What It Doesn't)

It is important to understand what this paper is not saying. The authors are not claiming they found a new planet, a new type of star, or a secret message from aliens. They are not saying the scrambled images are "wrong" in a physics sense; they are just showing that the scrambled images lack the connectedness of the real thing.

Instead, this is a methodological breakthrough. It proves that if you only look at the "energy recipe" of a space image, you miss the most important part: the shape. The real universe has a "coherence geometry" that survives even when you try to break it. The scrambled images might have the same average brightness, but they are like a pile of sand, while the real image is like a sandcastle. The sandcastle has a structure that the pile of sand, even with the same amount of sand, simply cannot mimic.

The authors call this "informational coherence geometry." It's a new way to look at space images that helps scientists find the "glue" that holds faint, weak signals together. It doesn't tell us why the glue is there (that's for astrophysicists to figure out), but it gives us a powerful new tool to find where the interesting, connected structures are hiding, even when they are too faint to see with the naked eye.

In short, the universe isn't just a bag of ingredients; it's a carefully constructed recipe. And thanks to this new method, we now have a better way to taste the difference between a real cosmic masterpiece and a random mix of the same parts.

Technical Summary: Informational Coherence Geometry in Astronomical Imaging

Problem Statement
Astronomical image analysis traditionally relies on spectral, statistical, and model-based tools to characterize faint structures and suppress noise. However, standard Fourier analysis separates a signal into amplitude and phase components. While the amplitude spectrum describes energy distribution across spatial frequencies, it fails to preserve the phase-dependent spatial organization that determines connectivity, anisotropy, and persistence. Consequently, two images with identical Fourier amplitude spectra can possess vastly different structural organizations. The paper addresses the question of whether real astronomical images contain connected coherence structures that are not reproduced by phase-randomized images (surrogates) that share the exact same Fourier amplitude spectrum.

Methodology
The authors propose an "informational coherence-geometry pipeline" inspired by Viscous Time Theory (VTT), though implemented entirely using reproducible, public operations. The workflow consists of the following steps:

Preprocessing: RGB images are converted to grayscale and normalized. Bright point sources, diffraction spikes, and annotation regions are masked out to create a fixed analysis mask ( $B$ ) applied identically to real and surrogate data.
Surrogate Generation: A phase-randomized surrogate is created by taking the Fourier transform of the image, preserving the original amplitude spectrum ( $A(k)$ ), and replacing the phase $\phi(k)$ with a random field $\psi(k)$ subject to Hermitian symmetry. This creates a null model where spectral energy is preserved, but spatial organization is destroyed.
Local Coherence Mapping: The structure tensor is computed for the normalized image at a smoothing scale $\sigma$ . Local coherence ( $C(x)$ ) is derived from the eigenvalues ( $\lambda_1 \ge \lambda_2$ ) of the structure tensor, representing directional persistence. High coherence indicates strong anisotropy, while low coherence indicates isotropy.
Connected-Component Analysis: Pixels exceeding a coherence threshold are identified. The algorithm decomposes these high-coherence regions into connected components and calculates the size of the largest cluster ( $S_{max}$ ).
Statistical Assessment: The $S_{max}$ of the real image is compared against an ensemble of 50 phase-randomized surrogates. Empirical significance ( $p_{emp}$ ) is calculated as $1/(M+1)$ if the real value exceeds all surrogates.

Key Results
The method was applied to two distinct James Webb Space Telescope (JWST) NIRCam targets: the COSMOS Legacy Field (a deep extragalactic field) and Cassiopeia A (a resolved supernova remnant).

COSMOS Legacy Field:
- The real image produced a largest connected high-coherence cluster of 1,358 pixels.
- The ensemble of 50 surrogates produced a maximum largest cluster of 223 pixels.
- The real image had a lower mean background coherence (0.3686) than the surrogates (0.4783), yet formed significantly larger connected structures.
- The real-to-surrogate ratio for the largest cluster ranged between 8.0 and 10.6 across nine different threshold combinations, demonstrating robustness.
- Empirical significance: $p \approx 0.0196$ .
Cassiopeia A:
- The real image produced a largest cluster of 780 pixels.
- The surrogate ensemble maximum was 285 pixels.
- The real-to-surrogate ratio was 2.74 (relative to the max surrogate) and 5.53 (relative to the mean).
- Empirical significance: $p \approx 0.0196$ .

In both cases, the real images formed fewer but substantially larger connected high-coherence regions compared to the fragmented, smaller clusters found in the phase-randomized surrogates.

Key Contributions
The paper's primary contribution is methodological rather than astrophysical. It demonstrates that:

Coherence Topology is a Distinct Descriptor: Spatial organization (connectivity) contains information not preserved by Fourier amplitude statistics alone.
Mean Coherence is Insufficient: Surrogates can exhibit higher average local coherence than real images, yet fail to form extended structures. Therefore, connected-component analysis is necessary to detect phase-organized morphology.
Phase Sensitivity in Weak Signals: The pipeline successfully distinguishes real phase-organized structures from spectral surrogates in both deep-field and high-contrast remnant scenarios.

Significance and Claims
The authors explicitly state that these results do not identify new astrophysical objects, establish new cosmological structures, or prove the physical identity of the detected regions. The detected coherence regions may reflect astrophysical morphology, unresolved diffuse emission, correlated background structure, or residual processing artifacts.

Instead, the paper claims that coherence topology is a useful phase-sensitive descriptor for weak-signal astronomical imaging. It serves as a complementary, exploratory post-processing tool that can prioritize regions for further inspection. The method provides a reproducible way to test whether an image contains connected geometrical organization absent from a spectrally matched null model, offering a sharper descriptor of phase-organized structure than mean coherence or spectral analysis alone.

The Great Cosmic Scramble

The Big Discovery: Real vs. Random

What This Means (and What It Doesn't)

Technical Summary: Informational Coherence Geometry in Astronomical Imaging

More like this