Pattern Matching in Google Search Engine
Pattern matching in Google search is an aspect where Google automatically matches images with the extant library of objects which the images are related with using the specific keywords. In Google images, a user can be able to upload an image and Google will turn back an image that resembles the one uploaded. Google has a basic image grid that contains basic shapes such triangle, square, and circle. One could explore all images that are indexed on google through clicking on any shape. This act results into similar shapes being obtained. For example by clicking on a black and white circle, one may obtain an orange or sun in turn. You may then go on browsing images and getting similar ones in this manner.
In another instance, a user could upload the portrait of a celebrity through Google you -tube or video. The result will be a list of videos portraying the celebrity. For instance, one can upload an image of Ford Harrison, and what may be obtained in return is Ford’s trailer to Indiana Jones. Image matching in Google search is also applied on Google maps. In this perspective, one may click a house and what may come in return are images depicting similar houses. This is workable in both chains and architecture. For instance, if one clicks on the MacDonald’s, Google will indicate the location of all the other MacDonald’s in the nearby surroundings. Another feature in this perspective is that you can upload an image of a house which is unknown for the purpose of research and you will certainly find the location of this house. It does this by returning an image that is the best match street view of the uploaded house.
Through Google Ad Sense, Google is able to analyze the page colors that are dominant for the purpose of adjusting the colors in the Ad Sense according to the user’s preferences. It is also possible for the web owner to indicate colors that are dissimilar from one another but can effectively fit into the color scheme in accordance to the particular design formulas stored by Google. Also through Ad Sense, Google is able to analyze the combination of colors that can be able to obtain a high click rate of click thrues. This information can then be transmitted to webmasters in making them optimize their advertisements.
Through Google’s spam filter, Google is able to determine manually a category of sites that are spam seeds and evaluate if they are visually common. In the case when a particular design is found to match the common spam designs, then the score of its spams could be increased and when a particular threshold is reached, the site will be automatically banned.
Through Google Translator, people could upload a text image, which is then automatically translated into a selected language by use of character optical optimization. This is significant especially in regard to situations where there is only a text image. For instance, when the image text is a Chinese logo and one is not able to translate the characters in Chinese logo. Similarly, through Google translator, one could apply the CCR to any web URL which one is translating. This means that even sites that employ heavy graphics could as well be translated.
Also through Picas Web Albums, parts of an image could be tagged and Google would be able to indicate sections that are related in different images. For example, one could choose an iPod as an image, inform Picasa that it is an “iPod”and the images will be associated with the images 412, 132 and 130 which include this object with a keyword, “iPod” for a faster searching.