“My dog Oliver has always been quite vocal, and recently I decided that his thoughts and comments needed to be shared with the world,” Henry Conklin wrote on his blog this week.
Conklin didn’t want to just manually tweet every single bark. Instead, he built a kind of bark detector based on the Raspberry Pi, an inexpensive computer popular with hackers and tinkerers. The detector captures each bark, and then uses machine learning to differentiate actual barks from any other kind of noise.
Conklin then connected his bark detector to Twitter, where it now sends out tweets whenever Oliver hears the mailman, or finds any other reason to start barking.
At the moment, Oliver’s Twitter account just tweets out random barking phrases, but Conklin promises to tweak his technology and build what he calls a “bark-to-text translator” that will “be more accurate to Oliver’s actual voice.”
May we suggest Pixar’s “Up” for inspiration?