Remember TinEye? It’s a image search engine that allows users to input images as search query and get results based on the input image. The user uploads a picture or pastes the URL to a publicly available image on the web and TinEye returns links to web resources that host the same image. TinEye is an excellent tool to find sources of images without attribution. Google is working on a similar concept.
Based on a technology that debuted with Google Goggles, a mobile application that matches pictures snapped by your phone camera to those stored in Google's web-based image database, search-by-image lets users search images with images from the desktop.
To search using an image, go to images.google.com and just put your picture in the search box, by clicking on the camera icon in the search box and uploading a photo from your computer. You can also paste the URL of an image from the web, drag and drop pictures from webpages or your computer into the search box. To perform a one-click image search, you can also download the Chrome or Firefox extensions.
Search by Image then returns the best results for images that have related content on the web. You’re more likely to get relevant results for distinctive landmarks or paintings than you will for more unique photos like your kids. In addition to getting relevant results about your image, you can also find visually similar images or the same image in different sizes or resolutions. This is unlike TinEye, that returns only results that are identical to the input image.
The service does not do facial recognition per se. But there may be cases where it can identify images that involve faces, according to Google Fellow Amit Singhal.
If you can’t find the camera icon, don’t worry because the feature will be rolled out globally across Google Images over the next few days.
[via Inside Search]