Feb. 4, 2008 -- Dog owners may think they can recognize their pet's bark. But a new study shows that a computer is far better at identifying a canine's woofs, arfs, and ruffs than a human.
The computer, which relies on artificial intelligence and machine learning methods, could provide a new tool for automatically analyzing large amounts of data typically recorded during behavioral studies.
"Animals communicate very different emotional states by very different vocal sequences," said Csaba Molnár, an ethologist from Eötvös Loránd University in Budapest, Hungary.
"You could use this method to analyze other vocal signals, such as bird songs, or categorize video recordings. You could even use it to categorize human speech and dialogues."
The software, which was developed by collaborators at the Sony Computer Science Laboratory in Paris, France, analyzed more than 6,000 barks from 14 Hungarian sheepdogs known as Mudis.
The researchers recorded barks in six different behavioral categories: "stranger" (where an unknown person appeared at the house in the absence of the owner); "fight" (where a trainer encourages the dog to bark aggressively); "walk" (where the owner prepares for a walk outside); "alone" (where the owner tied the dog to a tree and walked out of sight); "ball" (where the owner held a favorite toy); and "play" (where the owner played a familiar game with the dog).
After the barks were collected, they were transferred to a computer and digitized. Artificial intelligence software is then used in a two-stage process to first learn the more than 100 acoustic features of the different kinds of barks and then use that knowledge to code, classify and evaluate them.
In one experiment, the software correctly classified the barks in 43 percent of cases. "Fight" and "stranger" barks were the easiest to recognize, while "play" barks were more difficult. When matched against a human's ability to do the same, the computer's success was about the same.
In a second experiment, the software correctly recognized the barks of 52 percent of individual dogs. A human's ability to do this was much lower, which suggests that there are acoustic differences to dog woofs that humans cannot discern.
According to veterinarian and applied animal behaviorist, Sophia Yin, animal vocalizations that sound similar on a cursory listening are very different upon closer analysis. In 2004, Yin and colleague Brenda McCowan published a similar study in the Journal of Applied Animal Behavior.
However, the computer they used employed statistical methods and not artificial intelligence to crunch the data.
"It might be that the method they're using is a faster way to analyze. Because we did statistical analysis, it took a long time. We had to keep running the machines overnight to get all of those barks analyzed," said Yin.
However the information is gleaned, discerning dog barks can give owners better information for dealing with the animal's behavior.
"People might just think, 'Oh he's barking again, we have to use an electronic collar on him,'" said Yin.
But if they can understand the context for the bark, they can find a better solution.
For future work, Molnár and his team plan to use the artificial intelligence software to compare the barks of various working dogs, such as sheep dogs and hunting dogs. A difference could provide some insight into the effect of artificial selection on the vocal communication of the dogs, he said.
Related Links:
our sites
video
mobile
shop
stay connected
corporate
