Last week, Google secured Indian patent for image identification wherein the images of a named entity in the search engine will be identified via the face-recognition method. The Search Engine will identify the named person, retrieve images tagged/associated with the named entity, and carry out face-detection using an algorithm.
The patent application for image identification was submitted by Google in 2009 at Mumbai. The application was approved on 16 April 2018 by the Assistant Controller of Patents, Designs and Trade Marks. So, how does the image identification process work? When the name of a person is typed into the Google search engine with regards to a query, the search engine will identify images tagged with the person’s name and display it in the results along with the requested content.
Google has filed a patent document that states that the image identification method is made foolproof by identifying minimum one representative face image of the searched entity and with that representative face image, more images of the entity can be identified. In order to ensure the search query gets appropriate results, the images have to be tagged with the person’s name and appear in metadata, etc. However, a large collection of images on the web that is taken from videos are not tagged. Therefore, those images may not appear in the search results.
Why is this image identification method necessary? Search queries about famous people/entities are popular on Google. Therefore, it is important to ensure the results contain the right images of the searched entity. Google’s image identification method via the face-recognition process will ensure the images displayed in the results are that of the searched entity.