Google has launched a new tool called 'Swirl', which will help out users to find images in an easier way. The tool automatically groups identical images into categories presented on results pages. "Whatever Image Swirl does is - it automatically organizes image search results into groups and sub-groups. It is not just the face, the color, the visual features of the images; we look inside the image and ask over how they relate to each other on a pixel level," said Aparna Chennapragada, the Google Product Manager, to AFP.Image Swirl expands on technologies developed for similar images as well as Picasa Face Recognition to discern how images must be grouped together and build hierarchies out of these groups. Each and every thumbnail on the initial results page represents an algorithmically-determined representative group of images with related appearance and meaning. These are not just the most relevant images – they’re the most relevant groups of images.
Image Swirl at present works for more than 200,000 queries and we plan to include more queries in the future. Available queries will auto-complete as you begin to type in the search box, like Google Suggest.
"As humans, we can easily tell which images must be grouped together by looking at them. The actual question is can we train computer algorithms to identify common features in images and then categorize them. It is not perfect. We may possibly have mistakes, but we hope to improve this system as we move forward," added Google Software Engineer Yushi Jing, who has been working on Image Swirl's underlying technology.
People may dabble with Image Swirl at Google Labs, and feedback will be used to develop the tool with the ultimate objective being to one day incorporate it in the California Internet giant's public search engine. "We’re committed to how we look at and understand visual information and organize it for our users," added Chennapragada. "This is the holy grail of computer vision."
Google launched image search in 2001 and later on added a 'similar images' feature to hone-in picture hunts.
No comments:
Post a Comment