An Impression-Word-Giving System for Images Based on Colors

Kazuhiko SHIRANITA, Kenichiro HAYASHI, Akifumi OTSUBO
2004 Transactions of the Japan Society of Mechanical Engineers Series C  
Saga 849-0932, JAPAN, 1.Introduction Printed matter is produced by directly using materials and photographs given by clients, or images and photographs which designers create based on the impression words and the colors given by the clients. In the manufacturing process of printed matter, it is important for designers to select images and photographs, that match with a client's reguest from a database of images and photographs. Until now, designers have repeatly
more » ... ve repeatly performed the selection of images and photographs which match with the client's request based on their extensive experience. However, it is very difficult for many designers to select images and photographs that will satisfy client's request, since there are many kind of images and photographs available in the image database. Generally, the selection process is time-consuming and laborious, and thus is inefficient. If a new system through which designers and clients smoothly understand each other based on impression words and colors were developed, the system would be useful in the manufacturing process because the work of selecting images and photographs would become efficient. Therefore, the development of an impression-word-giving system has long been expected for supporting communication of ideas between designers and clients. There have already been reported several studies on the giving of impression words for : We describe a method of giving impression words for images based on colors. 160 impression words Abstract are prepared, and colors which are strongly related to each of the 160 impression words are registered into the system in advance. In this system, we extract main colors of images based on the results of clustering in Lab color space. The number of main colors is determined when the sum of pixels of the main colors is larger than 85 % of all the pixels in the image and the colors with the larger of pixels are determined in order. We propose a distance between the main colors and the registered colors. Using the distance, we calculate a minimum distance and an average distance every 160 impression words. By the comparison of the minimum distance and the average distance of each impression word, the system determines suitable impression words for the image. Experimental results show that the proposed method is effective.
doi:10.1299/kikaic.70.192 fatcat:sqamuou74zgy3abphdnivmycvq