Artificial intelligence improves first image of a black hole

appeared in 2019 first image of a real black holeIt once again proves the work put forward by the General Theory of Relativity, published by the physicist Albert Einstein in 1915. Until then, the image was showing the shadow of the black hole with a ‘blurry’ appearance, so the scientists used a machine learning technique (machine learning) to improve the quality of the photo..

In a study published in the scientific journal This Astrophysical Journal LettersOn Thursday the 13th, scientists released an enhanced image of the black hole, perfected by the PRIMO (Principal Component Interferometric Modeling) machine learning technique. The same processing data as the previous photo was used, but the resolution was improved through image reconstruction algorithms.

“I affectionately call the previous image the ‘hairy orange donut’ and refer to this image as the incredibly unappetizing ‘skinny bun’. We also discussed the equally unappetizing ‘diet bun’.” Lia Medeiros, an astrophysicist at the Institute for Advanced Study in Princeton and lead author of the study, told Reuters in a message.

First image of black hole

The image shows the supermassive black hole at the center of galaxy Messier 87 (M87), located about 55 million light-years from Earth., confirming many physics predictions on the subject – it was first observed in 2017, but the explanation came two years later. The researchers were able to capture the image in partnership with the Event Horizon Telescope (EHT) project, a global network of radio telescopes capable of observing objects in space.

While in the first photo it was possible to visualize an orange ring of hot gas and the dark silhouette of the black hole, now the image appears to be a little more focused and showing a thinner ring of gas. In a video published by the study team, it is possible to observe that the machine learning technique improves the image.

The research was led by four scientist members of the EHT project; Medeiros says the team plans to use the enhancement technique to perfect the second image of a black hole—besides being the second, it’s the only other image of a celestial image. entity of this nature.