Image search goes to an even higher level. Meet Google Lens
The ability to find information on the Internet about something that can only be found in graphics has been possible on Google for almost 20 years. This year, Google Lens was introduced, which allows you to be even more precise about what you are looking for and get almost 100% accurate results. The feature can be found in the Google Photos app on smartphones, as well as in the standard photo search engine. Its innovation lies in the ability to outline / mark a specific element with accurate bullets. The mechanism of operation of Google Lens is shown below.
In the Google Photos app on the smartphone, we selected a photo that captured the insect we would like to identify. In order for the application to start analyzing photographs, press the icon marked with a cross in the screenshot.
The white dots in the screenshot below mean that the function has started working and is now looking for characteristic points in the photo. The process of examining the graphics takes a few seconds, during which the dots – reference points for the mechanism – move quickly, processing information. After a while we get the finished result. In our case, Google Lens has detected that the photo contains the plant Klarkia splendid. However, we meant an insect, so we suggest to the mechanism the appropriate area of the photo by selecting its shape or holding the screen on a given element – the insect is small, so just holding it worked, so that the application understands that we want it to repeat the search.
A larger circle with a white outline appeared in the photo, after which Google Lens suggested to us what the object we were selecting could be. A honey bee and… a dove was proposed, which the creature in the photo undoubtedly is.
The mechanism is also great for finding furniture – which was presented on the Google Poland blog – clothes or specific places.
A tool operating on similar principles is currently testing Unamo – Internet monitoring platform. It will be intended for agencies and brands, and photos and graphics will be analyzed for the presence of logotypes.
Now, the monitoring of specific phrases related to the brand will also apply to images. This is a useful convenience not only in the context of examining the presence of mentions of our company in social media and the Internet, but also in the case of spying on competitors. Served on a platter, we will have at least her idea for visual identification, which says a lot about the entire communication strategy. For now, the mechanism is in the beta phase, and the only thing we can do is provide contact details so that we will receive a notification about its launch in due time and see for ourselves how AI can still surprise us.