Jul 22, 2015 10:56 AM EDT
'Sketh-a-Net' Computer Program Can Identify Hand-Drawings Better than Humans
A team of scientists developed a computer program that could correctly identify three out of every four subjects of sketches.
In their study, the researchers compared the success rate of Sketch-a-Net, 74.9 percent, with that of a person, 73.1 percent, according to a Queen Mary University of London (QMUL) press release. The program's development is coming at a time where sketching is converging with computing due to the advancement of mobile devices with touch screens.
The program could allow a new method of Internet searching, which would come in handy when one does not know the name of something, but has a clear image in their mind. It could also be integrated into law enforcement by matching sketches to mug shots or images of suspects.
"It's exciting that our computer program can solve the task even better than humans can. Sketches are an interesting area to study because they have been used since pre-historic times for communication and now, with the increase in use of touchscreens, they are becoming a much more common communication tool again," study co-author Timothy Hospedales, a lecturer at the QMUL School of Electronic Engineering and Computer Science, said in the release. "This could really have a huge impact for areas such as police forensics, touchscreen use and image retrieval, and ultimately will help us get to the bottom of visual understanding."
The British Machine Vision Conference accepted the research, which was co-authored with four other researchers.
Join the Conversation