Let me begin this article with the case of the cat. I was quite surprised upon reading an article from NYTimes.com about how Google researchers came up with identifying a cat using images. Take note, it took them 16, 000 images in order to identify what a cat is.
We normally search something online using keywords but we can also already search the web by using images. Usually it’s the other way around wherein we search the web for some images.
Functions of search by image
Normally we can use this to check where the photo was taken, what’s it all about, and some other basic info about the content of the photo itself. However, this technology play several important functions which goes beyond checking what the image is all about.
According to TinEye, reverse image lookup can help you find the source of an image, research and track where the image appeared online, find and grab higher resolution versions of the image, and locate the websites that use your image.
What to use?
TinEye is simply the biggest reverse image search site now. It has 2.17 billion images in its database. The images are being crawled by TinEye and at the same time the site has also many contributors. It looks for exact match of your photo.
For Google, the reverse image search technology works through computer vision techniques. This simply matches the image that you searched for with their indexed images as well as with other collections. And they already have extensions for Chrome and Firefox.
There are still other options if you want to use reverse search. Here’s a list from comptalks.com:
- Byo Image Search (color similarity)
- Gazpopa (color and shape similarity)
- RevIMG (shape, dimensions, & colors)
- Multi service image search
- IM2GPS (location where the image was taken)
- SnapTell (perfect for shopping)
Some of these offer interesting features like locating where the photo was taken. Advanced search algorithms are also used which involve identification through color pallete uniformity (like the one used by Byo Image Search).
Google Search By Image gets revamped
Earlier this month, Google updated its reverse image search tool. Knowledge Graph is already incorporated in the results when you search for images. Of course, there are many related info that appear on the results page.
It also became smarter in its best guesses. An article in searchengineland.com, shows a classic example – the flower. Previously, when an image of a flower is searched for, Google only identifies the image as flower. With the update, it already attempts to identify which type of flower is in the image.
Is it really relevant?
In my own opinion, this technology is great. However, there are images that are not uploaded yet. This only works for online photos or images that have a copy somewhere in the web. So there is a high chance of you not getting any info about some old personal photos.
But overall, this technology is amazing because you can locate an image anywhere in the web. Moreover, there are different approaches being used, so you have other options in case the exact match approach doesn’t work.
Reverse image search has many practical uses. With the increasing number of photos in the databases used by the TinEye and Google, it is so much easy to track your own photos and to get info about an image.
So start checking your photos now. Maybe someone is illegally using it.