In this paper, we describe a comparative experimental study of sign language learning characteristics by hearing people between two types of still image materials: hand-drawn illustrations, and images synthesized by actual sign language video. The experimental results show following tendencies: (i) illustrated images are effective in learning handshapes, (ii) synthesized images are effective in learning sentences and facial expressions.

View full abstract

This article contributes to build a method and case study to link the natural movement and display effects of the user emotional in automatic cartoon generation using natural user input interface (NUI) like Kinect. As a suggested method, “KinEmotion” has posture evaluation functions which is configured with contribution ratios for each joint (bone in Kinematics) and the similarity ratio can be obtained by inner product with a target posture. The output of posture evaluation function can be assigned to some pictures which can used to express emotions respectively. The proposed method is not limited to particular users because of it is based on angular evaluation. Interaction designer can configure player interaction with a scripting and culled process can be 2.6 times simplified faster than past algorithm AccuMotion. The applied demonstration system “Manga Generartor” and its future applications and development are also reported with results of scripting abstraction in actual student project as a case study. The system and constants had been tested in some public demonstration and permanent exhibition.

View full abstract

We propose a non-photorealistic rendering method for generating ripple-pattern-like images from input images. In the proposed method, inverse sobel filter is applied to an input image for transforming it to a ripple-pattern-like image. In this case, we show how to generate the ripple patterns in the vertical, horizontal, and diagonal directions. To validate the effectiveness of the proposed method, we investigate the change in the ripple-pattern-like images on the image of Lena by changing the values of the parameters. Lastly, we evaluate the appearance of these images subjectively.

View full abstract