This paper gives an overview on an intelligent avatar communication system using Korean, Chinese and Japanese sign-languages to overcome the linguistic barrier between different languages. First of all, to achieve real-time communication, an intelligent communication method based on a client-server architecture and a sign-language translation method are considered. Also, to generate realistic gesture images, CG animation techniques and emotional expression methods are reviewed. Finally, to overcome the limitation based on different avatar models, a new communication message format using a FBML (facial body markup language) is investigated. Experimental results show a possibility that the methods could be used for intelligent sign-language communications between avatars of Korean, Japanese and Chinese on the Internet.

This paper proposes a novel real-time non-verbal communication system from natural language instruction by introducing an artificial intelligence method into the networked virtual environment (NVE). We extract semantic information as an interlingua from the input text by natural language processing, and then transmit this semantic feature extraction (SFE), which actually is a parameterized action representation, to the 3-D articulated humanoid models prepared in each client in remote locations. Once the SFE is received, the virtual human will be animated by the synthesized SFE. Experiments between Japanese sign language and Chinese sign language show this system makes the real-time animations of avatars available for the participants when chatting with each other, not just based on text or predefined gesture icons, so the communication is more natural. This proposed system is suitable for sign language distance training as well.

Watermarking is a technique embedding hidden information in multimedia data imperceptibly, such as image, video, audio and text data. Generally an original image is transformed and coded watermark data is embedded in frequency domain watermarking models. We propose a new image data embedding method using Fresnel transform. A watermark image is Fresnel-transformed and the values of difference between coefficients of a transformed pattern and their average are embedded into an original image. Our watermarking model allows flexibility in hiding the watermark. It is possible to get many embedding patterns from a single watermark image by using various distance parameters with the Fresnel transform. All kinds of image models, such as shapes, letters and photos, can be used as watermark data. Watermarking experiments have shown the validity of our watermarking technique and its robustness in resisting attacks.

Watermarking is a technique embedding hidden multimedia information imperceptibly, such as images and sounds. Generally original image is transformed and coded watermark data is embedded in the frequency domain watermarking models. In this paper, we propose a new data embedding method using Fresnel transform. A watermark image is Fresnel-transformed and the intensity of transformed pattern is embedded into original image. Our watermarking model has the flexibility in data hiding. It is possible to embed watermark in many different observable planes with various distance parameters of Fresnel transform and all kinds of image models such as shape, letter and photo can be used as watermark data. Using Fresnel transform with various distance parameters, it is possible to get many channels for embedding. So our method can recover the weak point having only one transform plane in DFT and DCT domain methods. The watermarking experiments were conducted to show the validity of the proposed method, and the results show that our method can provide multi frequency domains for embedding watermark and have robustness against filtering and geometric transformation.

The sign language can be used as a communication means between ordinary people who have no common language. In order to get over the linguistic barrier, we had developed a 2D model-based sign-language communication system between Korea and Japan on the Internet. From the study, we discovered that there are some problems to be solved for natural animation and real-time transmission in the communication system. In this paper, we employ a 3D character model for the stereoscopic gestures of the sign-language animation, where we utilize CG-animation techniques in order to generate unartificial gestures. For real-time communication, we make use of an intelligent communication method on the client-server architecture. Using the 2D and 3D models, we implement the communication system with Visual C++ 5.0 and Open Inventor on the Windows platforms and experiment on the comparison of their availabilities. Experimental results show that both systems have some properties in their own ways and could be used for non-verbal communication between different languages beyond the linguistic barrier.

We propose a chatting system that uses avatar language to communicate across the linguistic barriers. We developed a CG system for generating CG animations of words from avatar language. A communication experiment was performed using a Japanese communications satellite to confirm the validity of the developed chatting system.

We propose a sign language communication between different languages such as Japanese-Korean and Japanese-Portuguese using computer graphics (CG) animation of sign language based on the intelligent image communication method. For this purpose sign language animation is produced using gesture or text data expressing sign language. In the production process of CG animation of sign language, MATLAB and LIFO language are used, where MATLAB is useful for three-dimensional signal processing of gestures and for displaying animation of sign language. On the other hand the LIFO language, which is a descendant of the LISP and FORTH language families, is developed and used to produce live CG animations, resulting in a high-speed interactive system of designing and displaying sign language animations. A simple experiment was conducted to translate Japanese sign language into Korean and Portuguese sign languages using the developed CG animation system.

We propose a method to index baseball telecast for content-based video retrieval. In our approach, assumptions specific to baseball telecast are used to index video-recordings. In, the first stage of the system, we detect domain specific scene in a baseball video based-on image similarity. These scene, called basic scene in the paper, are the shots which include a single pitching in each. After extracting these scenes, we spot the exact location of pitching and batting action using continuous dynamic programming matching for fixed areas in the image. If the batter swing the bat, we determine the end point of the play from the camera view after batting to recognize the batting result. We also recognize the caption to verify/confirm the recognition result. We stored summarized version of the telecast with these indexes to form a video-database. To validate the approach, we parsed 53 throws in a baseball game using the system, and the percentages of correct spotting for pitching and batting are 96% and 89%, respectively. The validity of the indexing is examined in the report.

We are developing an intelligent presentation system. This type of system can handle live images of documents and drawings. In the system, documents are input using a computer-controlled camera; it can actively decide zoom parameters and glance directions of the camera in order to capture the whole document image in sufficient resolution. Scanned images are mosaiced into a large image. Then, links between document data and electronic data can be constructed by document analysis. In this research, we describe techniques in order to realize this type of system.

A new morphing technique is proposed, where two two-dimensional (2-D) models are Fresnel-transformed and intensities of the transformed patterns are superposed and clipped, resulting in a blending pattern of two models. Since the Fresnel transform depends upon the distance parameter between the model and the Fresnel-transformed planes, the morphing process from one model to another is observed by changing the distance parameter. The morphing experiment was conducted using 2-D models, such as letters and simple figures by performing Fresnel transforms with 32/spl times/32 sampling points. Discussion is done by comparing the proposed morphing technique with a morphing technique using Fourier transform, where model spectra are superposed in the spectrum plane using a weighting factor to blend the spectra of both models proportional to the weighting factor, whereas phase functions are multiplied on the Fourier transforms of models in the Fresnel transform method.

A situation-based selective video recording system for a short term memory aid is proposed. The system selectively records the images of events which happened in one's daily life based on sensor observation of the user situation. The user wears sensors to monitor the environment and the action of the user himself. Selected scenes viewed with a head-mounted camera are recorded with situation information. This paper reports the system overview and some field experiments.

This paper proposes a method of recognising Japanese sign language indirectly using a colored glove which has multiple color patches on its surface. Japanese sign language has finger alphabets and words which include a hand motion. In the proposed method, the finger alphabets are estimated from three parameters: the color combination of the visible patches, the dispersion of the visible patches, and the hand direction. The color combination is represented by the mean of the color histogram of the visible patches in an input image. Words which include a hand motion are estimated from the transition of characteristic hand structures which are detected from an image sequence by changes in the color combination of the visible patches. The experimental results show the validity of the proposed method.

In many OCR systems, character segmentation is a necessary preprocessing step for character recognition. It is an important step because incorrectly segmented characters are not likely to be correctly recognized. The most difficult case in character segmentation is cursive scripts. Uygur character is a cursive script. This paper presents the problem of segmenting the Uygur characters in various fonts and size in printed scripts. The technique for the segmentation is presented as following: line separation, word separation, segmenting the word into isolated characters consists of the two step's algorithms, topological segmentation, and quasi-topological segmentation. Topological segmentation is based on tracing the outer contour of a given word. Quasi-topological segmentation is based on the decision to section a character on a combination of feature-extraction and character-width measurements. Our approach relies on the feature of characters and fonts and profile models.

Embedded computers (EC) have been used widely, however, embedded computer software is often difficult to develop for the lack of a suitable debugging environment. With the traditional in-circuit debugging method, it is difficult to analyze the reason for errors; also it is impossible to debug a software before the implementation of the associated hardware; moreover, it costs both money and time to create the hardware, so that it can hardly keep up with the ever-changing requirements of a great number of ECs. To tackle the problem, we have researched a prototyping method for a simulation debugging environment (SDE) of embedded computer software. Our purpose is to establish an environment to construct SDE and use the software SDE to substitute the traditional in-circuit debugging environment. By the method, users can construct a SDE at a higher speed and lower cost, the development efficiency for embedded computer software will also be greatly increased.

We propose a qualitative analysis of group behaviour based on multi-scale region analysis. In the algorithm, histogram backprojection images of an input image sequence are analyzed with multi-scale filters to extract a hierarchical representation of the group structure in the sequence. The dynamics of the hierarchical description is interpreted from the temporal aspect of qualitative relations among groups. Occlusions are also detected with transition rules. A soccer game image sequence is chosen as an example of the analysis.>

A method fo transmitting sign language images by an intelligent communication is proposed. In this method text data is transmitted to synthesize sign language images at the receiving end. Dictionaries for coding and decoding sign language images are developed, and a stick figure model is adopted to draw the sign language images using computer graphics (CG). Joint angles are given, and bending fingers and arms are described by the connecting vectors. Based on the evaluation of the synthesized sign language images evaluated by handicapped persons, surface and solid models are used to display 3-D sign language images, and an animation with controlled hand motion speed is synthesized to improve the understanding of communication by sign language. An experiment using a Japanese communications satellite was conducted to verify the effectiveness of the proposed communication method.>

In our previous paper (1990), we discussed methods of expressing the composition knowledge of Chinese flower and bird paintings and proposed the dividing and composing method which could support computer painting based on the composition rules. In this paper, we attempt to apply this dividing and composing method to the various frames of paper (paintings using paper of different shapes) of Chinese flower and bird paintings. First, we analyzed the composition rules and described the expression method of composition knowledge in different frames of paper. Then, we defined the characteristic volume of painting elements, the frame centroid, the number of divisions and the division displacement to extract a dividing model. This procedure made possible the composition formation of various frames of paper.>

One of the problems in sensor integration is how to design the integration strategy for the given task. The authors deal with model-based object recognition from uncertain geometric observations using uncertain object models. First, they decompose the recognition problem into a hierarchy of statistically well-defined subproblems depending on sensor uncertainties and model uncertainties. A recognition algorithm based on this approach is developed. Second, a method to preserve the consistency under model uncertainties is discussed. It is shown that information loss can be avoided by adding dummy variables to parameters in the integration. Finally, applications of the proposed method to 2D object recognition are demonstrated.>

A new technique to process images of under-snow radar with a displaying system of three-dimensional information is proposed and an experiment to display three-dimensional radar images is done with the constructed under-snow radar system. Since the principle of the proposed radar system is holographic imaging radar, a large IC memory system of 256 K byte, which corresponds to 128 × 128 × 128 boxcels, is constructed to store hologram data and reconstructed images, where each memory cell of 128 × 128 pixcels has 7 bits data corresponding to 128 gray levels of depth information. An experiment to display radar images of the burried objects in snow is done using X-band microwave. Improvement of resolution of images of under-snow radar is also discussed using the constructed dispaying system.

The design algorithm of a differential group programmable logic array (DGPLA) to generate the precise binary logarithm function is suggested. It can reach an optimal condition such that the number of bits in a PLA is minimized, while the error is still kept as small as possible. Thus, the space in the PLA is saved, estimated at only 15.94 percent of the space for a readonly memory (ROM) counterpart.