Much experience is necessary to play a musical instrument. In general, it is difficult for beginners to practice a musical instrument. Therefore, there are many supporting systems for playing a musical instrument. In the case of a keyboard, we can play it by just tapping the key shown by the appropriate light indicator. However, you use both hands and feet at the same time when playing the drums. Therefore, we cannot play it by the method for supporting keyboard performance. In this paper, we propose a motion generation method for supporting drum performance using MIDI files. This method helps the beginner to understand the motion of drum performance. Some experimental results show that the proposed method is effective for the practice of drum performance.
A method of generating the textures of dirty floors is presented. First, the method computes the position and speed of each walker by simulating the behavior of the crowd. Then, footmarks caused by soil, dust, black grime on the floor and scratches caused by small stones are displayed. Footmarks and scratches are represented by adding shoe sole patterns and hair lines to the floor image, respectively. The amounts of dust and grime are increased with time. However, in the area stepped upon, their amounts are reduced by a fixed percentage. By implementing this, footmarks and scratches appearing at the center of the corridor are displayed. Dust and grime on the wall are displayed, too. In addition, dirty floors are cleaned up at regular intervals. We assumed daily and full cleaning. By repeated cleaning, the images of floors of old but well maintained buildings, on which only a lot of scratches are viewed, can be generated.
We propose a method for modeling human skin including macules and show the rendering results. This method is intended especially for applications in dermatology and cosmetics. Our model is based on volumetric representation. It consists of nine layers classified by the difference in optical properties. The surface normals and the number of voxels of each layer are derived from measurement data of the real skin surface. Each voxel has a type of layer that indicates the values of the optical parameters. The density distributions of the melanin in the macules are defined by a gray-scale image. We use photon mapping for rendering to consider multiple scattering and spectral attenuation. Rendered images with or without macules can successfully reproduce the color, and the spectral reflectance is also shown.
We propose a novel system, called CosmicAI, for providing an intuitive and efficient way to synthesize 2D sky images. The system utilizes a collection of high-quality proprietary sky part images for backgrounds, clouds, and the moon, and allows the users to generate their own sky images effectively through attribute-based retrieval of these sky parts images and their flexible layout and natural composition based on a new cloud extraction algorithm and an original filter for partly-overlapping pixels. Its one click search and active composition play an important role for realizing rapid image synthesis. Moreover, CosmicAI is capable of collecting via the Internet and searching for composite works. This function promotes group works by giving incentives to users with unclear design motives for generating a wide variety of sky images.
We introduce a non-photorealistic rendering technique for generating line art color images from photographs. Line art color images are pictures that consist of only solid color lines. In this letter, line art color images are generated using the staircasing effect of a bilateral filter. To evaluate the performance of our method for generating line art color images, we conduct experiments using some example images.
We have proposed a non-photorealistic rendering method using the staircasing effect of a bilateral filter for generating moire picture-like images from photographs. This moire pattern is uniquely determined by the nature of the bilateral filter and original images. Therefore, we propose a method for synthesizing another moire pattern that exerts some controls to generate moire picture-like images. We validate the effectiveness of the proposed method by conducting experiments using the image of Lenna and other example images.
In Japanese digital terrestrial television broadcasting (DTTB), the identification of multi-path is important to improve its reception characteristics. To enhance the resolution of the delay profile, connecting continuous multiple channels and the extended frequency bandwidth are effective. However, the guard band is existing between each channel and the transfer function is not obtained directly. The phases of each channel are not synchronized because the each channel is broadcasted independently. In this paper, a method to connect multiple channels is proposed. Interpolation of the guard band by using the liner adaptive predictor based on the RLS algorithm and compensation of phase between adjacent channels are performed. As a result, it is confirmed that resolution enhancement of the delay profile is possible. It is possible to improve the time resolution of the delay profile to below 0.024[μs].