映像情報メディア学会誌

抄録を表示する抄録を非表示にする

Individual creative design with personal computers has become common. However, at least in the animation design process, two problems remain: cumbersome setting tasks for design parameters and few environments in which users can simultaneously handle parameter setting and animation preview. We thus propose a novel interface for line animation design using relative changes in non-verbal voice information. In the interface, a user associates voice parameters (volume or pitch) and design parameters by uttering in time with an animation. Then, design parameters can be adjusted by uttering again in accordance with the relative change in voice. In a preliminary experiment, we confirmed that there are universal correspondences between the direction of change in design parameters and the direction of change in non-verbal voice. In addition, from users monitoring with the system, we verified the validity of non-verbal utterances and revealed the characteristics of voice input.

抄録全体を表示

PDF形式でダウンロード (4921K)

Projectors have recently been widely used not only in theaters but also in accessible environments such as living rooms with non-flat or non-white walls. In these environments, image projection can be quite difficult due to the geometrical and photometrical distortions. We have developed an adaptive image projection method to overcome such dynamic environmental challenges. In order to achieve highly accurate and interactive compensation, we approximate a projector's response function on the basis of changes to the target reflectance property and environmental light, including the effect of indirect reflected light. We also use a depth sensor to measure geometrical changes in real-time to correct geometrical distortions. We implemented the proposed method with a ProCams system and a GPU and discuss the results for actual environmental changes in an ordinary room.

抄録全体を表示

We propose a coding method that reduces inter-channel correlation of 4:4:4 chroma format in high-resolution images more than HDTV. The proposed method adaptively conducts color conversion on the singular value decomposition using local decoded pixels in H.264/MPEG-4 AVC. The conversion coefficients are calculated from pixels that are similar to the target block on the coding parameters of H.264, including intra prediction mode and motion vector. Experimental results show that the proposed method has 0.7 dB better coding efficiency than the conventional method and is about two times faster.

抄録全体を表示

Hand-printing styles can be preferable to normal printing in some documents, such as those for elementary school students. Although many font sets have hand-printing appearances, figures with hand-printing appearances are still hard to handle because such figures are not explicitly supported by most documentation systems like Office Word. This paper describes a scaling method for AutoShapes tools that handle figures with hand-printing appearances. The proposed scaling method avoids unnatural exaggerations of hand-drawn fluctuations during scaling operations. By using this method, an AutoShapes tool was developed as an add-in for Microsoft Word. In experiments, the subjects made positive responses about the developed prototype.

抄録全体を表示

J-STAGEへの登録はこちら（無料）