In order to improve communication environments among the hearing impaired and the hearing people, the authors are developing a system translating Japanese into Japanese Sign Language (JSL) and vice versa. Evaluation tests by the hearing impaired were carried out on JSL animation of binocular vision and plane vision. Results showed that the animation of binocular vision has obtained high ratio in comprehensibility and some important factors for designing JSL animation were indicated to develop the sign animation.
The aim of this paper is to develop a system for bidirectional translation between Japanese and Japanese Sign Language (JSL) for assisting the hearing-impaired. This system can display the translated JSL into Japanese and Japanese into JSL by sign animation which a model avatar signs. Various displays, from small screens such as a mobile telephone to large ones of a projection type, are available to a sign animation. We have to find display condition which is hearing-impaired well understand sign animation. Experiments were carried out under 45 display conditions and hearing-impaired read sign animation on screen. Results show that sign can be read with correct answer rate 90%, when the size was over 10x8 cm and the resolution was 225x180 pixels.
Non-Photorealistic Rendering on two dimensional digital images by the computer image processing technique is shown in this paper. Fundamental purpose of this research is to realize the system to make the images with various artistic expressions and build specialized image processing software for Non-Photorealistic Rendering. Non-Photorealistic Effect is artistically highlighting effect, for example pencil drawing style, watercolor painting style, and oil painting style. Here, we take particular note of pencil drawing style. Starting with discussion and deliberation of existing processes, we developed an image generation software of the pencil drawing style, moreover, we discussed about Non-Photorealistic Rendering with the partial image processing method to be applied to the face-recognition technology.
The optimal hardware implementation method of vector graphics for embedded systems is described. The feature of our method is the reuse of one-line buffer, without sorting of intersecting point. Moreover, speeding up of in-out test and clipping process made high-speed and high-quality drawing of vector graphics on the compact hardware. We have implemented one of our architecture to FPGA evaluation board, and OpenVGTM application successfully demonstrated on the hardware.
This paper describes the extraction method of the contours of an auricular cartilage.First, the section of the upheaval portion of an auricular cartilage and a cave-in portion is modeled with a Gauss function, and the characteristics required for threshold processing is extracted. Furthermore, the contour of a cartilage is extracted by the method of Canny using the extracted characteristics and it evaluates at the end.
Due to develop Internet, there are a lot of information around us. Then the image search system which picks up necessary things in such information is expected. In this paper, we propose the method to search “human's face image”. We determine similarity which emphasizes color of human's skin by using color vector quantization and visualize the search result by SOM. Experimental result shows that this method improves search accuracy of faces to non-emphasis. And visualization by SOM mapping can help user search image.
This paper deals with shot change detection and classification for broadcast soccer video. To extract interesting scenes in soccer games such as goal scenes, it is essential to know the movements of playersand a ball. Since broadcast soccer video is composed of many shots from several cameras, we need to extract shots which capture a wide area of the ground (called ground camera shots) for estimating their movements. We therefore develop a method of detecting shot changes, which can work even for those with a special effect such as fade in, and that of classifying shots into several ground camera shots and others. Experimental results with about 25-minute video show the effectiveness of the method.
It is important to divide video data semantically for retrieving it efficiently. It is desirable to divide it in consideration of camera motions. Because of this, video creators often put their minds into their camera motions. Therefore, in this paper, we propose a method of detecting camera motions and dividing video data. In camera motion analyses, there are two problems: camera vibration and foreground object motion. In our method, to avoid the effect of camera vibration, we used spatio-temporal images for detecting camera motions. Also, to avoid the effect of foreground object motions, spatio-temporal images in consideration of composition of video data.
Recently, Task Support Services in ubiquitous environment attract more attention. Appropriate model of tasks and estimation of the situation of tasks from sensor information are necessary for such services. We are researching about Cooking Navigation System synchronizing with action which can recognize user's cooking progress and show the appropriate recipe corresponding to the cooking situation. In this paper, we describe about the Modeling of Cooking Task for recognition system of cooking situation using events that is detected by sensors.
In this paper, we propose a novel example based marker free motion capture method. First, volume data of actor's body and its skeleton are obtained from multiple camera views at each frame through visual hull based method and thinning process. Then, examining the structure of the skeleton, we can acquire an attributed graph and comparing the acquired graph and the graphs in the Model Graph Database, correspondences between body parts and skeleton parts are obtained. Finally, we estimate the position of the joints examining the curvature of the skeleton. The experiment result shows the validity our approach.
A cellular system is used in communication between a cellular phone and a base station. In the traditional cellular system, a system administrator designs the cell size to accommodate as many nodes as possible. In general, it has been pointed out that reduction of cell size increases communication capacity per unit area. This also leads to decrease of nodes in the vicinity of border between cells. In such a case, a base station can avoid radio interference among its adjacent base stations by appropriately controlling its transmission power (spatial partition among cells). In this paper, we first propose a simple algorithm to improve connecting probability based on an existing radio interference model. Then, we introduce a geometrical analysis of the connecting probability according to the proposed algorithm. Through simulation experiments, we first show the validity of analysis. Furthermore, simulation results show that the power control can improve the connecting probability up to 80% compared with the traditional method.
3D human model reconstruction using visual cone intersection from the images captured by multi cameras has been researched a lot, but there is a limit in the accuracy of the shape estimation because the method reconstructs from the silhouette images. Even if you increase the number of cameras to improve the accuracy, the problems, that unsmooth surfaces appear in the shape data and the normal of the surface is lost because of the characteristic of the method, will not be solved. So, in this research, we try to refine to process reshaping reconstructed shape. We estimate the posture from a captured human shape data, and create a standard human model to fit standard human model parts that was prepared beforehand. Next, we reshape and fit between human shape data and standard human model using method of alignment with reshape parameters. By this technique, we can get an accurate and refined 3D human model that maintains original human's posture and characteristic of his body.
Ubiquitous environment, where people can use communication resources everywhere, is expected to bring novel communication services. Position information of communication nodes plays important role, at the design of such ubiquitous communication systems. From position information, we can make graphical expressions, namely images. T, relation between ubiquitous communication-design architecture and image-related architecture are broadly compared, using following examples. (1) application of ubiquitous/wireless communication resources, (2) traffic analysis of cellular systems, utilizing probability density function of communication nodes, (3) capacity analysis of Ad Hoc networks, utilizing probability density function of communication nodes, (4) topology control of mobile communication networks.