Image-based search is one of Google’s most exciting, but still lesser-known and used features. In this article, I’ll show you in detail how you can use Google Image Search on any device, desktop or smartphone.
When searching for images on the web, the traditional solution is to type a term (keyword) that interests us into Google search and then browse through the result by clicking on the “Images” tab.
For image-based search, the image itself will be the “keyword”, the starting point, and the search result will also be an image. This is why this feature is also called “reverse” search.
Google Image Search – What’s Good for?
Image search can be a very useful tool, for example, when
- you want to trace the origin of an image because, say, you want to ask permission from the owner to use it
- you want to know where the illustrations, photos and other images posted on your website have been used on the Internet
- or just looking for a new image for your own website
- you may be wondering which images are indexed by Google related to your domain address.
Google’s image recognition algorithm analyzes images on the web from a variety of perspectives. Able to interpret the shapes, colors, backgrounds and locations shown in the images, to recognize that there is a human in the given image, or rather an animal, possibly some object or product.
Examines the metadata attached to the images (alternative texts, file names, descriptions and captions), also analyzes the size and resolution of the image file. It also takes into account the context of the images and the website on which the photo, drawing or even logo appeared.
The algorithm therefore draws conclusions from a huge heap of information and gives search results.
For the sake of example, I downloaded a picture from a BBC article about swallowing dolphins swimming in New Zealand soon for tourists because it’s bad for animals. The picture shows people swimming with dolphins. I renamed the file – to make the algorithm a little harder – and then started a search on Google image search at images.google.com.
Image above is the result. In short, Google shoots very precisely what the image is about. It pops up a search term that matches an image (“new zealand south island dolphins”) and ranks first in two websites that offer dolphin swimming to tourists. Show another stack of images that show the same theme, and then list the pages where that image already exists. First comes the BBC article from which I took the picture.
Maybe that’s how it seems
Google Image Search is a very advanced feature that is really worth using every day.
Incidentally, image-based search is not Google’s exclusive invention, the pioneer in the field was the also free TinEye app.
Types of Google Image Search
There are basically three ways to use Google Image Search:
- The easiest way to use image-based search is to grab an image from the web, right-click, and then select Image Search with Google (Search Google for Image).
- Image search by URL. To do so, go to images.google.com , click the photo icon below the Google logo, and choose to paste the URL of the image.
- Search by uploaded image. We can also upload images to images.google.com and start searching based on them. In the same way, click on the small photo icon and then select Upload Image.
Google officially stores images uploaded to Image Search for only 7 days. However, in the opinion of many, this does not mean that a copy of the images does not remain part of the search giant’s database in some form.
What results will you get?
You will see these in the image search results list:
- the size of the image
- the search term that matches that image
- the pages that Google links to the image
- a bunch of similar, related images
- all websites that already have the image you used to start the search with.
In addition, if the given image depicts a well-known topic, we will get some additional basic information on the page in a separate box. For example, if we search for a photo of San Lorenzo Cathedral in Genoa, we get this result:
You can further filter image results based on a number of criteria. By default, the image size and publication date are the two configurable filters, but depending on the topic, you can also sort by color, type (GIF, ClipArt, etc.), usage rights.
For example, you can easily search for snow-only, black-and-white images from the last 24 hours:
Image based search and SEO
There are three ways to use image-based image search to optimize your website for search engine optimization.
- Link building. If you use unique images or infographics on your website, search for it using Google’s image search feature. If you see your images appearing on other websites, write to them to put up your link. You already have a free link.
- Competitor research when creating new entries. When creating content for a new keyword, such as a blog post, it’s a good idea to see how many and what images your competitors (= who are ranked by Google for that keyword) are using (such as unique or free photos of their own or reused from elsewhere). From this, you can deduce what visual content Google rewards for each topic.
- We want to know what images Google associates with our website. Use the site: something.com command in the search box to list all the images that Google has indexed in connection with our website.
Image-based phone search
If you use the Chrome browser on your smartphone, simple image search works the same as on your desktop computer. To do this, you need to hold your finger on the image and then select Search Google for this image from the menu that appears.
However, image search based on URL or uploaded image will not go through images.google.com on your smartphone. You can use the Google Lens application for this.
Face recognition by image: possible?
If you’ve gotten this far by reading this article, you’re not at all surprised that Google can identify people based on the faces in a photo.
Of course, this only works really well when it comes to a known person. For example, in the case of the Prime Minister of India, we get a pinpoint result.
Although Google’s facial recognition capacity is growing (and scary), we probably won’t get too many meaningful results right now if we upload a photo of an average Indian resident for search.
In the future, of course, this may even change radically.