Companies send out many types of information to customers through media in the form of flyers, catalogues, websites, etc. However, a customer receiving information sent this way must decide whether to read it. Therefore, designing a document according to customer tastes to increase the possibility of the document actually being read is important.
In order to identify what kinds of images and design elements in media are preferred by people who are the target of communication, Fuji Xerox is conducting research on a "Taste Model" and a "taste and image evaluation method" called "Media Image GAP Finding."
"Media Image GAP Finding" is a tool used to visualize the gap between a design created by the sender of information (e.g., a company) and the taste of the recipient of that information (e.g., the customer) to determine such design elements as color according to customer tastes.
This concept of designing documents according to customer tastes is based on the "Taste Model," which categorizes the diverse tastes of people into 40 types.
(Fig. 1 shows the "Taste Model" mapped on an image scale.Note1)
The method of determining the taste and image of a flyer using "Media Image GAP Finding" is now being introduced. First, a flyer is read in bit map format, with its color distribution being extracted by using the hue and tone analyzing method (Fig. 2).
Then, color correction is performed according to the type of color distribution (categorized by whether the colors used are similar colors, contrasting colors such as red and white, etc.), and the center of gravity coordinates are calculated by mapping the color distribution on the image scale (Fig. 3).
Lastly, words at the coordinates of the center of gravity are extracted to determine the impression that a flyer gives to customers, and the "Taste Model" to which the flyer most likely belongs (Fig. 4).
By using the "Taste Model," the image that target customers prefer can be determined (or estimated), and the image of media created for communication with the customers can be analyzed as described above. By mapping the results on a single image scale, whether the image of created media matches the taste of customers can be determined.