“With very little training, people can spot real supernovae candidates,” said astrophysicist Arfon Smith of the University of Oxford. “They’re as good as any professional astronomer would expect to be.”

Supernovas are the bright bursts from exploding stars that have reached the end of their hydrogen-burning lives. These dramatic flares can briefly outshine entire galaxies, but their brilliance quickly fades. A supernova appears to Earth-based observers as an unexpectedly bright star that was not so bright the last time you looked. But other objects, like approaching asteroids and variable stars, can trigger false alarms.

Discriminating software can throw out about 90 percent of the fakes. But the last 10 percent had to be sifted through by humans, who, until about a year ago, were “a small but tiring team of [about eight] PhD students” sorting through up to 5,000 images a night, Smith said.

The new project, Galaxy Zoo Supernovae, takes the work to the masses. Galaxy Zoo Supernovae is an offshoot of the citizen science powerhouse Galaxy Zoo, which began asking at-home volunteers to identify galaxy shapes in 2007 and has since branched out to the moon, solar storms and even historical weather data.

In the supernova project, amateur enthusiasts look through images from the ongoing Palomar Transient Factory survey to sort the dead stars from the living.

Each participant is shown three images of the same region of the sky: one image that the survey telescope just captured an hour ago, a shot from days or weeks earlier, and an image where the “before” shot has been subtracted out and only the supernova candidate remains.

The citizen scientist is then asked a series of three or four questions designed to rule out supernova impostors.

More than 2,500 volunteers sorted through nearly 14,000 supernova candidates between April and July 2010, Smith and colleagues report in a paper submitted to the Monthly Notices of the Royal Astronomical Society. Citizen scientists correctly identified 93 percent of the brightening objects, with no false positives.

“At the moment, it’s something only humans can do,” Smith said. “People’s brain cycles are really valuable.”

The crowd of citizen scientists was also faster at classifying the supernova candidates than the small team of professional astronomers had been.

“All of our candidates get classified within 20 minutes every day,” Smith said.

That’s exciting for supernova experts, who want to catch and follow up on explosions as quickly as possible.

“The first few hours or days after a supernova explosion reveal information that is crucial to furthering our understanding of what kinds of stars explode and how the explosions occur,” said supernova expert Alex Filippenko of the University of California, Berkeley. “Imagine studying an adolescent or an adult; it would help to have photographs, videos, and audios of that person as a baby.”

Ultimately, though, the Galaxy Zoo crew would like to cut out the human middlemen. They team is using the data collected from Galaxy Zoo Supernovae to train the next generation of supernova identification algorithms.

“It’s a feedback loop: classifications from people can help make algorithms stronger in the future,” Smith said. “I think if we were still running this project in 3 years time, I would be disappointed.”

Image: Galaxy Zoo Supernovae

See Also:

• NASA Needs You: 6 Ways to Help an Astronomer
• Milky Way May Fizzle Out Sooner Than Expected
• Rare Dead Quasar Found in Nearby Galaxy
• Galaxies Discovered in Their Awkward Teen Phase
• Citizen Science: Trawl World War I Navy Records for Weather Data

Follow us on Twitter @astrolisa and @wiredscience, and on Facebook.

Authors: Lisa Grossman

to know more click here