Google Labs focused more on online picture searching, on Tuesday, with a ‘Swirl’ tool that groups images of the same category automatically on search result pages.
“What Image Swirl does is it really organises automatically image search results into groups and sub-groups,” Aparna Chennapragada, product manager of Google, told AFP.
“It's not just the face, the colour, the visual features of the images; we look inside the image and ask how they relate to each other on a pixel level,” he added.
‘Swirl’ uses facial recognition software, metadata and algorithms to differentiate places, people, structures and time of day of the digital images.
For instance, the keyword ‘Washington’, inputted to the Image Swirl, generates result pages with images grouped into subsets, such as former President George Washington, the White House, and even movie star, Denzel Washington.
Clicking on one of the subsets opens up a window to additional ‘sub-clusters’ of images grouped by factors like time of day, angle and even whether a picture is a panorama or a close-up.
Online search for images has conventionally presented result pages with thumbnail images that individuals search through separately.
“As humans we can easily tell which images should be grouped together by looking at them,” stated Yushi Jing, software engineer of Google, who worked on the Image Swirl technology.
“The real question is can we train computer algorithms to identify common features in images and then categorise them. It's not perfect. We may have mistakes, but we hope to improve this system as we move forward,” he added.
“What Image Swirl does is it really organises automatically image search results into groups and sub-groups,” Aparna Chennapragada, product manager of Google, told AFP.
“It's not just the face, the colour, the visual features of the images; we look inside the image and ask how they relate to each other on a pixel level,” he added.
‘Swirl’ uses facial recognition software, metadata and algorithms to differentiate places, people, structures and time of day of the digital images.
For instance, the keyword ‘Washington’, inputted to the Image Swirl, generates result pages with images grouped into subsets, such as former President George Washington, the White House, and even movie star, Denzel Washington.
Clicking on one of the subsets opens up a window to additional ‘sub-clusters’ of images grouped by factors like time of day, angle and even whether a picture is a panorama or a close-up.
Online search for images has conventionally presented result pages with thumbnail images that individuals search through separately.
“As humans we can easily tell which images should be grouped together by looking at them,” stated Yushi Jing, software engineer of Google, who worked on the Image Swirl technology.
“The real question is can we train computer algorithms to identify common features in images and then categorise them. It's not perfect. We may have mistakes, but we hope to improve this system as we move forward,” he added.
RSS Feed