Search Box

Monday, July 27, 2015

Computer Recognition of Hand Drawings (Sketch-a-Net)

New Computer Program Recognizes Sketches Better Than Humans

Catherine Griffin | July 22, 2015

A new computer program may be better at recognizing sketches than you. Scientists have created a program, called Net, that's capable of correctly identifying the subject of sketches 74.9 percent of the time as opposed to humans, which only are correct 73.1 percent of the time.
Like Us on Facebook.
Sketches are intuitive to humans, and have been used as a communication tool for thousands of years. However, recognizing free-hand sketches can be difficult since they're abstract, varied, and consist of black and white lines rather than colored pixels like a photo.


<more at; related links: (AI triumphs at Pictionary-like sketch recognition task. July 22, 2015) and (Sketch-a-Net that Beats Humans. [Abstract:  We propose a multi-scale multi-channel deep neural network framework that, for the first time, yields sketch recognition performance surpassing that of humans. Our superior performance is a result of explicitly embedding the unique characteristics of sketches in our model: (i) a network architecture designed for sketch rather than natural photo statistics, (ii) a multi-channel generalisation that encodes sequential ordering in the sketching process, and (iii) a multi-scale network ensemble with joint Bayesian fusion that accounts for the different levels of abstraction exhibited in free-hand sketches. We show that state-of-the-art deep networks specifically engineered for photos of natural objects fail to perform well on sketch recognition, regardless whether they are trained using photo or sketch. Our network on the other hand not only delivers the best performance on the largest human sketch dataset to date, but also is small in size making efficient training possible using just CPUs]; further: (New computer program is first to recognise sketches more accurately than a human. July 24, 2015)>

No comments:

Post a Comment