Now search images using images, not keywords

BANGALORE, INDIA: Have you ever tried to find a photograph online to see where is the image posted on the Internet, to check out who else had re-posted the same image or to scout for similar photos?

Reverse search engine is the answer! Reverse image search is a different kind of search where one does not provide keywords, instead provides an image and the search engine searches for images of similar kind.

The so-called search engines like Google Image search have one thing in common -the keyword comes first, words are used to search for images. However, reverse search uses an algorithm which can identify objects in an image and return matching results from its database. Such search engines enable you find same images of better resolutions, modified versions of the image being searched and to see comments and feedback about the particular image.

Let's have a look at some of the search engines that are reverse in nature and use images to find out similar images, rather than keywords.

TinEye

The image search engine called TinEye use image identification technology rather than keywords or metadata. It is free to use for non-commercial searching.

Simply upload an image or provide link of the same online to find out where it came from, how it is being used, modified or to find higher resolution versions. TinEye crawls the web for new images, and also accept contributions of complete online image collections.

When you submit an image to be searched, TinEye creates a unique digital signature or 'fingerprint' for it, then compares this fingerprint with other images in its index to retrieve perfect matches. TinEye does not typically find similar images, it finds exact matches including those that have been cropped and modified. See the screenshot.

The reverse image search engine also has a add-on for Firefox, Chrome, IE, Safari, or Opera. It can be a great tool to find stolen or un-attributed images on the Internet. We installed the plugin for Chrome and tried to search images using images and image URLs. TinEye smartly found out results that very well matched with images searched with.

{#PageBreak#}

Byo Image Search

Byo Image Search is a method of searching through a collection of images by visual attributes such as shape, color, texture, etc. The search engine does not rely on the use of keywords or any other metadata ?it relies on the appearance of the image. This search is based on the color palette uniformity primarily but does include theme based algorithm. They use a search algorithm called Piximilar visual. You can use either a photo that you upload, or a photo that's already online. You'll need to limit yourself to a 1MB size for photo. I randomly picked some images from my desktop to see how well it returns results. The search engine wasn't successful in giving me perfect matches like we got in TinEye, butshowed a list of pictures with the same color palette.

GazoPa

GazoPa is a similar image search service on the web. Users can search images from the web based on user's own photo, drawings, a photo already available online or even keywords.

GazoPa enables users to search for a similar image from characteristics such as a color or a shape extracted from an image itself.

GazoPa helps you find images that just simply can't be described by keywords.

{#PageBreak#}

SnapTell

SnapTell is slightly different and works just like the name implies. You snap an image and it tells you details of the image including its reviews, price, etc. All you need to use this is to simply take a picture of an item or its barcode. SnapTell then searches for all of information it can find regarding the image. Currently, this works with an image of books, CDs and video games.

Reverse image search engines are sure to be more reliable in time, though there is still room for improvement for such services. It is truly a remarkable technological advancement.