Quantcast
Channel: TextLinkBrokers Blog » Search Engine Optimization
Viewing all articles
Browse latest Browse all 5

Google Image Search: Faster, Smarter … Better?

0
0

google-image-search 8-29-12It’s no secret that Google is obsessed with smarter search – you only have to compare today’s search options with those of the original Google homepage to see how sophisticated the art of search has become. Google’s Holy Grail is understanding a user’s intent, and, with this objective in mind, text search has long featured varying degrees of personalization and autosuggestion. As of July 2012, Google’s Search by Image is heading down the same track.

Search by Image Appears

Image search has its roots in Google Goggles, a still-current mobile app that launched in 2009. Goggles allows users to take a picture with a phone camera and submit it directly to search, making it “a perfect answering machine for your visual questions.” In June 2011, Search by Image appeared, with “new techniques and functionality that optimize the experience for desktop,” including drag-and-drop query and one-click image-search browser extensions.

The limiting factor for image searches has always been the combination of processing power and data volume required for acceptable accuracy and speed. At the launch of Search by Image, Google claimed to have “billions of images” available for matching, and as of July 2012, a drag-and-drop search using a picture of the Statue of Liberty returns 1,760 results in just over three-quarters of a second.

Competitors Emerge

Google hasn’t had this playing field entirely to itself – Microsoft claims that seven percent of all searches on its Bing search engine are for images, and Bing’s Image Search feature has many of the bells and whistles championed by Google. Although Bing announced major image-search upgrades in June 2012, it still lacks drag-and-drop functionality, and its major selling point is its clean, minimalist interface with thumbnails “that shine.”

Google’s response was almost immediate – at the start of July, its Inside Search blog unveiled “smarter, more comprehensive Search by Image results.” Search algorithm improvements mean better “best guesses,” one of Google’s many techniques for telling you what it thinks you want to know. Once its algorithm decides what the image is likely to be, it identifies the subject and presents the user with information about it.

This is displayed as a Knowledge Graph, a feature recently added to Google’s text-search results pages. The new answer-based, or predictive, capability means that users will often locate the information that they want without leaving the results page, something that Google sees as moving firmly toward the grail.

Thanks to an expanded data index, Search by Image has an improved ability to locate other sites that contain the search image, presenting them as part of a composite results page. Although Google doesn’t say as much, the inference to be drawn is that previously, collating results to this level of detail would have been an unacceptably lengthy process.

The future

Predictive search is set to become the next tool for defining intent, and will doubtless become an increasingly essential element of all types of search – image-based or otherwise. Speaking in June 2011, Google’s Eric Schmidt told the All Things Digital conference, “We’re trying to move from answers that are link-based to answers that are algorithmically based, where we can actually compute the right answer.”

Yahoo, too, has adopted the “answers, not links” mantra, and although the jury is still out on its future as a leading search company, Yahoo engineers will be leading the development of “the next generation of search” as part of its contribution to the search alliance with Microsoft.

There are those who see the future of image search embodied in a tablet-like device, capable of interpreting speech, audio, video and text, that displays predictive results immediately capture of the relevant data is complete. Snap a shot of a tourist attraction, and instantly your tablet tells you what it is, where it is and anything else that it considers relevant about it. Hover over a foreign-language menu, and you have an instant translation – no questions asked.

Whatever your opinion of predictive search – and you can turn it down, if not off entirely – it will be increasingly evident as a key element of Search by Image in future. Just be careful where you point your tablet – you might not like what you learn.


Viewing all articles
Browse latest Browse all 5

Latest Images

Trending Articles





Latest Images