New technique validates historic black hole image
A small Australian team developed an independent imaging method that confirmed the landmark 2019 picture of the black hole inside galaxy M87.
Four scientists from CSIRO created advanced machine learning tools to analyze the original telescope data without knowing the expected result beforehand. This unique approach successfully reproduced the black hole shadow with almost identical size and shape.
The researchers applied elegant mathematical algorithms to remove calibration errors and biases that often affect images made from sparse measurements across the globe. Their method combines triplets of data instead of standard pairs to eliminate potential inaccuracies in the final reconstruction.
Dr. Samuel Lai compared the difficulty of capturing such distant objects to photographing a doughnut on the moon from a backyard. He said their model had no prior idea of what the black hole looked like during the analysis process.
Director Dr. Laurent Loinard stated that these independent studies are crucial for verifying results and strengthening community trust in scientific findings. The results prove the ring structure surrounding the central black hole is a genuine feature rather than an artifact.
“It is like standing in your backyard and trying to take a photo of a doughnut on the moon”

