There is a particular irony in the search for distant worlds: the instruments we build to see the faintest light in the universe can be blinded by the nearest thing in their field of view. A star is a million times brighter than the planet beside it, and that glare, for all the engineering devoted to blocking it, never quite disappears. The James Webb Space Telescope has spent years capturing images of distant planetary systems, and in at least one case, the data it returned contained a planet that no one could see. Now, a new technique has found it.
The planet is a giant world orbiting a star called 51 Eridani, about ninety-seven light-years from Earth. It had been discovered before by ground-based telescopes, but in Webb's observations, taken at a wavelength of 4.3 micrometers, it remained invisible—lost in the residual glare that leaks past the telescope's coronagraph, the instrument designed to block starlight . The planet's signal was there, buried in the data, waiting for a better way to read it.
That better way came from a team at Northwestern University. They developed a technique called FLIP—Physics-based Learning for Imaging Planets—which builds a digital twin of the telescope's optics, modeling how light travels through its mirrors and detectors, accounting for tiny imperfections and instabilities that shift the pattern of glare . Instead of relying on reference images of other stars to subtract the unwanted light, FLIP simulates the telescope's behavior with millions of adjustable parameters, reproducing the glare more accurately and removing it more completely . The improvement in sensitivity ranged from two to fifteen times, depending on the observation .
When the team applied FLIP to the 51 Eridani data, the planet emerged. It was not a new discovery in the sense of finding something unknown, but it was a demonstration of a capability: the ability to recover faint planets from archival data that conventional processing had failed to reveal. "Planets are a million times fainter than their stars," said Rodrigo Ferrer-Chavez, the study's lead author. "So, to image exoplanets, we need to remove that glare. That is a very tricky problem because even the tiniest changes in the telescope can affect how that glare appears" .
The significance extends beyond a single planet. Webb has observed dozens of stars in its search for exoplanets, and each observation contains data that existing techniques may not fully exploit. FLIP suggests that some of those datasets may hold planets that have not yet been identified, waiting for the right analytical tool . The technique relies on known physics rather than artificial intelligence, using optimization methods common in machine learning but grounded in the optical behavior of the telescope itself .
There are limits to what this means. FLIP does not make Webb more powerful in a hardware sense; it makes the data more legible. It will not find Earth-sized planets around Sun-like stars, which remain far beyond the reach of current direct imaging. But for giant planets in wide orbits, the kind that Webb can already detect in principle, it expands the range of what can be seen. The researchers suggest it could be applied to future observatories, including NASA's planned Habitable Worlds Observatory, which will attempt to image planets more like our own .
For now, the planet at 51 Eridani is visible again, recovered from the glare that had swallowed it. It is a gas giant, massive and distant from its star, not a world that anyone expects to harbor life. But its recovery is a small victory in the larger project of learning to see what is there, even when what is there is almost impossibly faint. The universe does not make it easy. The least we can do is build better eyes.
The images accompanying this article are AI-generated and do not depict the actual telescope, planet, or researchers.
Published by Banx Network. This article is part of the Banx decentralized media programme, powered by the BXE token on the XRP Ledger.






