Skip to main content
Search is a fundamental cornerstone of the internet.  Without an effective search engine we’d have a far more difficult time finding the information, products, and services that we look for every day.  The majority of web search focuses on text-based input with results shown in the form of links and descriptions.  Although this approach is highly effective for most people and situations, it is important to consider scenarios where alternative search techniques could be beneficial.  Let’s explore one alternative search solution – the use of images and graphics to help us find products, services, and websites on the internet. Visual search involves the use of photos and graphics to help find information.  Images can help with search in two ways: using images as the search criteria rather than text phrases, and using images to help filter search results down to the specific items of interest. Using images as such criteria involves sending a photograph to the search engine, which can then analyze the image based on characteristics such as color, shape, and size.  These characteristics are then matched against images in the search engine’s database, returning similar results.  Using an image as the search criteria is particularly useful when you might not have a clear understanding of what you’re searching for, or can’t easily translate your idea into a phrase of text.  For example, if you’re looking for a specific bottle of wine you might try a search phrase like “red wine with a duck on the label”.  With a standard text-based search engine, it might be necessary to try repeated guesses before finding the one that best matches your idea.  However, with visual search, you can simply upload a photograph and the system can tell you what the item is called. The technology behind searching with images has been in existence for many years, but it has only recently started to be integrated into well-known consumer products.  One of the most well-known is Amazon’s Flow mobile application, which lets you simply snap a photo with a mobile device, and Flow will automatically search the product database based on your photo.  This can help save time by eliminating the need to manually translate product names into text, as well as making hard-to-describe items easier to find. Another way in which graphics can help with search is with the display of search results.  Traditionally, search engines display results as lists of links and text.  Lists can be scanned by the eye fairly quickly, but it does take time to process each title and determine if the result is relevant.  By using images rather than lists of text, we may be able to improve the speed at which a person can find the item(s) of interest.  You can very quickly pick items out of a visual set based on color, size and shape because the eye is quickly drawn to familiar images.  This can lead to a significantly reduced length of time scanning the search result set.  This mirrors situations in the real world where a person needs to sift through a large number of items to find those of interest, such as when looking through a store shelf trying to find a specific item. There are a few different approaches to graphical web search engines.  Sites such as Oolone display the results of web searches as a set of website screenshots.  This could be very helpful if you know the color and layout of a webpage, but can’t remember exactly what it is called.

websites searches

Other sites such as Oskope display products in a visual manner.  This can help you locate familiar products based on size, shape, and color.  Oskope’s site lets you view the product selection of many popular eCommerce sites, such as DVD box covers.  Even though the text on these covers is usually too small to read, you can very quickly locate a specific item if you have an idea of what the cover looks like.  These search systems can help you find products when you know what they look like, but not what they are called.

 collection of dvds

Although visual search has been around for many years, it is starting to make its way into consumer applications, especially with the growth of mobile devices.  Visual search can offer interesting and attractive efficiencies over traditional text based input and output in many scenarios, better mirroring the natural ways in which people look for items.