The sharper image can help theorists better understand the physics of black holes, while the technology used to create it can be applied to other types of research, including the study of alien planets orbiting distant stars.

The newly processed image, published Thursday in the Astrophysical Journal Letters, used machine learning to fill in a lot of data missing from the original obtained by the Event Horizon Telescope. The EHT is not a single instrument, but rather a consortium of telescopes across the planet that collected data in a technique called very long baseline interferometry.

This method is very powerful, allowing astronomers to effectively mimic a telescope with a dish the size of Earth. But that telescope is full of holes, and the missing observational data poses a challenge to the astronomers.

Enter “PRIMO,” a machine-learning algorithm that was trained specifically on thousands of high-fidelity simulations of matter falling into black holes.

The new image captures the radiation emitted by superheated matter that gets whipped around the black hole as it falls in. That ring of light is about 2.6 times the diameter of the “event horizon,” the point of no return for infalling matter, said lead author Lia Medeiros, an astrophysicist at the Institute for Advanced Study in Princeton, N.J.

“The event horizon itself is not an observable feature. What we are seeing is what we call the black hole shadow,” Medeiros said.

Medeiros said the next target for enhanced imagery is Sagittarius A* (also known as Sgr A*, which is pronounced “sadge-ay-star”), the supermassive black hole at the center of our own galaxy that had its EHT portrait released last year.

Sheperd Doeleman, a Harvard astrophysicist and leader of the EHT collaboration, praised the creativity of the new approach, though he also noted that machine learning has some limitations.

PRIMO “is partially relying on the truth of these computer models to fill in the gaps in the EHT,” he said. “This lets it create images that are a little crisper, but whether the images are more accurate will require verification when we have more data.”

And the EHT is far from done. Doeleman said that his team is now doubling the number of telescopes across the planet that will observe black holes, enabling a broader range of radio frequencies. This will include small radio telescopes at optimized locations around the world. That effort will fill in many of the imaging holes and will enable production of high-resolution images, including movies.

“With more data,” he said, “we will soon all be watching true black hole ‘cinema.’”