In April 2019, a fuzzy orange ring appeared on screens around the world and changed science forever.
It was the first image of a black hole ever captured, a glowing halo of superheated gas surrounding the shadow of a monster 6.5 billion times the mass of our Sun, sitting at the centre of a galaxy 55 million light-years away.
Behind that image was an algorithm called CHIRP, short for Continuous High-resolution Image Reconstruction using Patch priors, developed by a then-PhD student at MIT named Dr. Katie Bouman.
Now a full professor at the California Institute of Technology, Bouman is pushing far beyond that first image.
Her lab is building the tools that could one day produce not just pictures of black holes, but real-time movies, capturing the swirling gas, magnetic fields, and flaring energy around these objects as they actually evolve, frame by frame.
She is also pioneering a technique called Orbital Black Hole Tomography, which would allow scientists to reconstruct a three-dimensional map of the region surrounding a black hole for the very first time.
The first image was a beginning.
What comes next is something no human has ever seen.
How a Graduate Student Helped Build an Earth-Sized Telescope
The story of that original image requires a brief trip back to 2013.
Bouman joined the Event Horizon Telescope project while still a graduate student at MIT’s Computer Science and Artificial Intelligence Laboratory.
The challenge the team faced was almost comically impossible on the surface.
A black hole, by definition, emits no light.
What you can image is the shadow a black hole casts against the glowing ring of gas and plasma orbiting just outside its event horizon, the boundary beyond which nothing, not even light, can escape.
But even the largest black holes appear vanishingly small from Earth.
According to the Event Horizon Telescope collaboration, imaging M87’s black hole required the same resolving power as reading a newspaper in New York from a café in Paris.
No single telescope on Earth could do that.
The EHT solution was to link eight radio observatories across the globe, from Hawaii to the South Pole, into a single virtual telescope roughly the size of the Earth itself.
But linking dishes scattered across continents generates enormous amounts of incomplete, noisy, corrupted data.
Bouman’s core contribution was developing the mathematical framework to turn that sparse, messy data into a trustworthy image.
CHIRP, the algorithm she led the development of, uses a technique called regularized maximum likelihood imaging, filling in the gaps between the telescope data by applying intelligent priors about what a realistic image looks like, without letting those assumptions bias the result too strongly.
To verify the approach was working, different imaging teams at EHT worked completely independently, each blind to the others’ results.
When the images converged on the same glowing ring, the team knew what they were seeing was real.
The Moment the World Saw the Image
April 10, 2019, was one of those rare days in science when the news stops ordinary people in their tracks.
The black hole, designated M87*, sits at the centre of the Messier 87 galaxy in the Virgo cluster.
As MIT News reported, more than 200 researchers from 80 institutions across the globe contributed to the result, which was published simultaneously in six papers in a special issue of The Astrophysical Journal Letters.
The image confirmed predictions made by Einstein’s general theory of relativity about how spacetime curves around a massive object.
It showed the shadow of the black hole, a dark region about two and a half times smaller than the event horizon itself, surrounded by the bright ring of lensed light from the accretion disk.
As the European Southern Observatory described it, the ring appeared brighter in the south, consistent with the direction of the gas orbiting the black hole at close to the speed of light.
The dimensions matched theoretical predictions with extraordinary precision.
The photo that circulated online was not from a single moment of observation.
It was reconstructed from data collected across four nights of observation in April 2017, collected across five petabytes of hard drives that had to be physically flown to processing centres because the internet was not fast enough to transmit the data.
Then in May 2022, the EHT team released a second landmark image: Sagittarius A*, the supermassive black hole at the centre of our own Milky Way galaxy, located approximately 27,000 light-years from Earth.
Bouman co-led the imaging team for this result as well.
Why Sgr A* Was Harder to Image Than M87*
Imaging the Milky Way’s own black hole turned out to be significantly more difficult than imaging M87*, even though it is much closer.
The reason comes down to size and speed.
M87* is 6.5 billion solar masses, enormous enough that gas takes days to weeks to complete one orbit around it.
Sagittarius A* is only about 4 million solar masses.
That sounds large by any earthly standard, but in black hole terms it is much smaller, and the gas surrounding it completes a full orbit in just a few minutes.
As the EHT collaboration explained, imaging Sgr A* was like trying to photograph a puppy chasing its tail at high speed.
The brightness and structure of the gas around the black hole changed significantly between individual observations.
The team had to develop entirely new computational methods to account for the rapidly evolving structure and average across thousands of possible images to produce a reliable result.
Bouman’s group at Caltech played a key role in developing those techniques, including tools that allow researchers to study how the image of Sgr A* evolves over the course of a night.
The Next Frontier: Moving from Photos to Movies
The images released in 2019 and 2022 were extraordinary achievements.
But to Bouman, they are also a starting point.
The two images showed time-averaged structures, static snapshots of environments that are in constant, violent motion.
What her lab is now working toward is something far more ambitious: capturing the dynamic evolution of a black hole in real time.
The next-generation Event Horizon Telescope, or ngEHT, is the project designed to make that possible.
Where the original EHT linked 11 radio antennas across the globe, the ngEHT will expand that network by adding approximately 10 new dishes at carefully chosen geographic locations, dramatically improving the coverage of the virtual telescope.
The US National Science Foundation awarded a $12.7 million grant to design the ngEHT, with the goal of enabling what the project describes as real-time movies of spacetime at the event horizon.
Harvard’s Center for Astrophysics, which is leading ngEHT antenna development, describes the ambition clearly: the goal is to study black holes through cinema, not photography.
These movies would capture the dynamics of gas spiralling into a black hole, the formation of relativistic jets, and the flickering flares of energy that erupt near the event horizon.
All of it happening at the boundary of the most extreme gravitational environments in the universe.
Orbital Black Hole Tomography: Seeing in Three Dimensions
Beyond movies, Bouman’s lab has introduced a concept that takes black hole imaging into an entirely new dimension, literally.
Every image captured by the EHT so far has been a two-dimensional projection, a flat representation of a three-dimensional structure surrounding a black hole.
Bouman’s group is now developing Orbital Black Hole Tomography, a technique that would use the known physics of how gas orbits a black hole to reconstruct the actual three-dimensional structure of the emitting region from those 2D projections.
As described in a colloquium at the University of Arizona’s Wyant College of Optical Sciences, this technique integrates known gravitational physics with a neural network representation to map the evolving gas emission around a black hole in 3D for the first time.
The approach draws on the same logical insight as medical CT scanning.
A CT scanner takes multiple 2D X-ray images from different angles around a patient and uses mathematical reconstruction to create a full 3D model of the internal anatomy.
Orbital Black Hole Tomography applies similar logic, but with the added complication that the source itself is moving and evolving as the observations are being made.
Research from Bouman’s group at Caltech, co-authored with researchers at Google and Princeton, has already demonstrated the approach on the M87* data, showing that 3D structure can be recovered from existing 2D observations when physics-based constraints are applied intelligently.
A 2024 paper in Nature Astronomy, involving Bouman’s group and collaborators at MIT Haystack Observatory, took this further by applying polarimetric tomography to a flaring event near Sagittarius A*, mapping how the polarised emission, which reveals magnetic field structure, evolved in three dimensions during the flare.
That represents the first time scientists have been able to map the evolving 3D magnetic field structure near the event horizon of our own galaxy’s black hole.
Why This Research Matters Beyond the Images
The scientific stakes behind Bouman’s work extend well beyond the photographs themselves.
Black holes are not just extreme curiosities.
They are the engines at the centre of most large galaxies, including our own, and the jets of energy they launch shape the formation and evolution of stars and galaxies across cosmic history.
Understanding how material falls into a black hole, how magnetic fields organise themselves near the event horizon, and how those fields channel energy into the relativistic jets that blast outward for thousands of light-years requires seeing the process directly.
As Universe Today reported in February 2026, the EHT collaboration recently released new high-resolution images of M87’s jet, tracing the stream of charged particles from the black hole’s edge outward.
These images begin to connect the structure of the black hole environment to the mechanism driving the jet, something no previous observation had been able to do directly.
The tools Bouman’s lab is developing, including score-based priors that use machine learning to inform the imaging process, are central to extracting this kind of detail from the data.
From Algorithms to Asteroids: A Career That Has Moved Fast
Bouman received her PhD from MIT in 2017 and joined Harvard’s Center for Astrophysics as a postdoctoral fellow before moving to Caltech in 2019 as an assistant professor.
She was promoted to associate professor in 2024 and to full professor in 2025, according to her Caltech faculty page.
She holds appointments across three Caltech departments: Computing and Mathematical Sciences, Electrical Engineering, and Astronomy, a spread that reflects how her work sits at the intersection of computer science and astrophysics.
Her awards include the Royal Photographic Society’s Progress Medal, the Electronic Imaging Scientist of the Year Award, and a share of the Breakthrough Prize in Fundamental Physics, which the EHT collaboration received in 2020.
In 2021, asteroid 291387 Katiebouman was named in her honour by the International Astronomical Union.
That last recognition carries a certain symmetry.
A woman who dedicated her career to imaging objects so distant and dense that light cannot escape from them now has a small piece of the solar system bearing her name, permanently in motion against a sky she has helped us understand more deeply.
The Computational Imaging Revolution
One of the broader impacts of Bouman’s work is how it demonstrates what computational imaging can achieve across fields.
The techniques developed for black hole imaging, extracting meaningful signals from sparse, noisy, incomplete data, have applications far beyond astrophysics.
In an interview with Caltech, Bouman described early conversations with colleagues in geophysics about applying similar methods to localise the origins of earthquake clusters, a problem that shares deep structural similarities with the imaging challenges of the EHT.
Her group has also worked on cloud tomography, reconstructing the three-dimensional structure and particle distribution of clouds from 2D images taken by aircraft or satellites, using the same mathematical ideas that power their black hole work.
The same has been explored in MRI imaging, where her group has published work on learning task-specific strategies for accelerated scanning.
Each of these applications involves the same core challenge: learning more than the raw data seems to offer, by encoding the physical structure of the problem into the mathematics of the reconstruction.
That is the insight at the heart of Bouman’s career.
Not more telescopes, not better cameras alone, but smarter algorithms that squeeze every last piece of information from the signals that exist.
What Comes Next
Bouman’s lab is currently working on several fronts simultaneously.
The ngEHT is progressing toward completion by the end of this decade, when it will provide the sharpest images achievable from Earth’s surface.
Alongside that, the Black Hole Explorer (BHEX), a proposed space-based instrument, would add an orbiting satellite to the ground array, extending the baseline beyond the diameter of the Earth and enabling detection of the photon ring, an extremely narrow ring of light formed by photons that have orbited the black hole multiple times before escaping.
The photon ring is predicted by general relativity with extraordinary precision, and detecting it would provide one of the most stringent tests of Einstein’s theory ever attempted.
The films, the 3D maps, and the photon ring all represent chapters in a story that began with that first blurry orange ring in 2019.
Each one requires solving problems that did not have solutions until someone like Bouman sat down to invent them.
A Reminder of What Curiosity Builds
There is a particular kind of satisfaction in realising that some of the most groundbreaking images in the history of science came from a graduate student who asked a simple question: what would you need to know about mathematics and signals to photograph something that emits no light?
The answer turned out to require years of work, a planet-sized telescope assembled from radio dishes on four continents, and an algorithm that could reconstruct truth from fragments.
And now the next question is already being asked.
Not just what does a black hole look like, but what does it look like in the next minute?
What does it look like in three dimensions?
What does it look like when it flares?
The images of M87* and Sagittarius A* that stopped the world were only the opening frame.
The movie is still being made.
Sources: Event Horizon Telescope | MIT News | Caltech faculty profile | Caltech CMS | ngEHT | Harvard CfA | Nature Astronomy 2024 | ESO | Universe Today
Insert this YouTube video at an appropriate place
