There are specific systems that enable you to search using image files. This article will introduce you to a few useful ways to do this.
At Google Images, you can search for image files using a plain-text query, or by either uploading an image from your computer or using an image URL from the web.
TinEye is another strong player on the image engine market. The company develops applications that recognise image files, and their aim is to make e.g. photographs searchable, not through descriptive data (keywords, metadata), but from their actual content. In their basic search function you can upload an image stored on your own computer, or save the URL of a image online. One of the main features of their search is that it gives you the chance to find out whether your image, or a modified version of it, has been used on another web page without your permission. Or, you can find out where the image has been used, and so on.
At this moment TinEye provides access to more than 40 billion images for analysis. Moreover, the the company also has an interesting commercial product portfolio, e.g. TinEye Alerts, MatchEngine, and WineEngine 🙂
This works on a similar principle to TinEye, gathering information about image files indexed by Google, Bing and Yandex. Note that the primary purpose of the service is to identify cases of copyright infringement in the use of images, and like TinEye there is a commercial product portfolio, but the simple reverse lookup is helpful for anyone to identify and find out more information about any image. So, for example, if I’d like to identity this:
I simply upload it to Image Raider and in a few seconds it returns the following results:
WolframAlpha Identification Project
Finally, take a look at this interesting WolframAlpha activity that is focused on image content intelligence. Upload your image and see what WolframAlpha thinks about it. The information isn’t always accurate or useful (for example, my mammoth image returned the result “person”, but then I also have the chance to submit a correct identification to help the AI behind the scenes get a little smarter), but this work is still in its early stages.
- There are systems that search for images based on a submitted photograph or other image file.
- You don’t need to know keywords or other metadata