EunJin Kim, a Ph.D. candidate of Color lab, has presented her recent work entitled, “Key Color Generation for Affective Multimedia Production: An Initial Method and Its Application” at 2016 ACM Multimedia Conference in Amsterdam, Netherlands. In the paper, the insights from designers’ color selection process were discussed and utilized to develop a key color generation algorithm. The algorithm creates a color by combining a genuine hue and a mood tone, and the result of a user test showed a satisfactory performance of the algorithm compared to designers’ selection. As a prospective application, the algorithm was embedded in a video editing system to create an attractive promotional video from user-generated contents.
Abstract
In this paper, we introduce a method that generates a key color to construct an aesthetic and affective harmony with visual content. Given an image and an affective term, our method creates a key color by combining a dominant hue of the image and a unique tone associated with the affective word. To match each affective term with a specific tone, we collected color themes from a crowd-sourced database and identified the most popular tone of color themes that are relevant to each affective term. The results of a user test showed that the method generates satisfactory key colors as much as designers do. Finally, as a prospective application, we employed our method to a promotional video editing prototype. Our method automatically generates a key color based on a frame of an input video and apply the color to a shape that delivers a promotional message. A second user study verifies that the video editing prototype with our method can effectively deliver the desired affective state with a satisfactory quality.