Google Images is a web-based product by Google for searching for images online. While it performs the same basic querying and result-fetching functions as Google’s flagship search engine, it's better understood as a specialized offshoot.
While Google Search produces web pages with text-based content by scanning text-based content directly, Google Images returns image media based on entered keywords, so its process looks a little different under the hood. The main factor in determining what images populate your results page is how closely search terms match image filenames. This, by itself, isn’t usually enough, so Google Images also relies on contextual information based on text on the same page as an image.
As a final ingredient, the algorithm leverages primitive machine learning, in which Google Images learns to associate certain images with one another to create clusters, to provide its reverse image search feature.
Once a search is submitted, the service returns a set of thumbnail images correlating to your keyword description.
At this point, users can access web pages containing a selected image, provided the website hosting the image allows this. If a website lets you view the page with the image, it may also let you directly access the image and open a page with just the image on it, essentially presenting the image’s individual resource-specific URL. Websites don’t always let you access the exact page with the image — sites which sell professional photography are an one example — but they do in many cases.