Global Sources
EE Times-India
Stay in touch with EE Times India
EE Times-India > Embedded

Microsoft, Google one-up humans in image recognition

Posted: 20 Feb 2015     Print Version  Bookmark and Share

Keywords:image recognition  deep learning algorithm  neural network 

If you once thought computers cannot beat humans at image recognition, Microsoft and Google will prove you wrong. Just recently, Microsoft has programmed the first computer to beat humans at image recognition, with a 4.94% error grabbing neural network compared with the human benchmark of 5.1%.

Five days after Microsoft's announcement, Google announced it had exceeded Microsoft by 0.04%.

The competition is fierce, with the ImageNet Large Scale Visual Recognition Challenge doing the judging for the 2015 championship on Dec. 17. Between now and then, expect to see a stream of papers claiming they have one-upped humans too.

ImageNet, with hundreds of object categories and millions of example images, has been running the competition since 2010. About 50 institutions are competing, but this is the first year that a computer will take the crown from the best human score. All contestants are using deep learning algorithms, which are all derived from various versions of artificial neural networks, which mimic the way the human brain works to varying degrees.

Image recognition

Most of the contestants freely provide papers describing their algorithm in great detail—in the spirit of open source without providing the exact code—explaining why their algorithm worked so well. Here, Microsoft revealed it was using deep convolutional neural networks with 30 weight layers. Google revealed its batch normalisation technique that keeps from saturating neurons during initialisation.

"In previous work, the neural units were hand-designed and fixed during training. In contrast, we make the units smarter by allowing them to take a more flexible form," Jian Sun, principal researcher for the Visual Computing Group, Microsoft Research Asia, told EE Times. "More importantly, the particular form of each unit is learned by end-to-end training. We observed that introducing smarter units can considerably improve the model."

When questioned further as to why their current neural network was able to take the crown as the first to beat the human experts, Sun responded by citing details of its deep learning algorithm, which usually initialises by training on 1.2 million training images. It then verifies on 50,000 validation images and finally applies what it learned to 100,000 test images in the main image database. Microsoft, however, took a slightly different tactic.

1 • 2 Next Page Last Page

Comment on "Microsoft, Google one-up humans in i..."
*  You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.


Go to top             Connect on Facebook      Follow us on Twitter      Follow us on Orkut

Back to Top