Physical beauty may lie in the eye of the beholder, but astronomers seem to have a different concept of beauty. Results for the 5th Interferometric Imaging Beauty Contest were announced earlier this summer and detailed in a paper posted to the arXiv yesterday. What kind of beauty contest is this, you ask?
Think less swimsuit modeling and more supergiant stars for starters. Simulating a technique known as interferometry (more on that later), the organizers tested astronomers on their ability to turn raw data into pictures. Using their choice of astronomical software, competitors had to create the most beautiful reconstruction of two test images.
The contestant who created the best match to the original "truth" image won the beauty contest. Now we can reduce beauty to a standard, quantifiable measure: calculating pixel-to-pixel differences. Stargazing truly is romantic.
When astronomers point their telescopes toward the night sky, faint, distant stars can be hard to see. Without high resolution images from these stars, scientists have great difficulty learning more about the star's size and mass. That's where interferometry can help.
Astronomers can overcome resolution problems with the help of multiple telescopes from different locations. Two or more distinct telescopes can focus on a single star and collect information about the star's emitted electromagnetic radiation. This may come in the form of radio waves, infrared or even waves visible to the human eye. Most commonly, interferometry has been used for radio waves.
Combining waves from multiple telescopes allows astronomers to ultimately convert these cosmic signals into a picture. Increasing the number of telescopes and the distance between the telescopes can boost the resolution of these images.
For the beauty contest, contestants had to construct images of two astronomical objects: a disc of gas around a newly formed star and a bloated supergiant star. The synthetic image of the infant star was created using a modeling program while the supergiant image was based on real science data.
After creating these test images, the organizers stripped them down into datasets like the ones acquired by telescope arrays -- the raw data that astronomers convert into images.
With these datasets, the competitors were required to duplicate the original images without being able to see them. The competitor with the lowest pixel disparity between his or her created images and the original "truth" images was crowned the winner.
Eight teams competed, and John Monnier from the University of Michigan won the contest. Personally, I prefer a more "Supreme Court" definition of beauty: I know it when I see it. And from my vantage point, astronomy is a beautiful science indeed.
If you want to keep up with Hyperspace, AKA Brian, you can follow him on Twitter.