We propose a system that, using video information, segments the mouth region from a face image and then detects the protrusion of the tongue from inside the oral cavity. Initially, under the assumption that the mouth is closed, we detect both mouth corners. We use a set of specifically oriented Gabor filters for enhancing horizontal features corresponding to the shadow existing between the upper and lower lips. After applying the Hough line detector, the extremes of the line that was found are regarded as the mouth corners. Detection rate for mouth corner localization is 85.33%. These points are then input to a mouth appearance model which fits a mouth contour to the image. By segmenting its bounding box we obtain a mouth template. Next, considering the symmetric nature of the mouth, we divide the template into right and left halves. Thus, our system makes use of three templates. We track the mouth in the following frames using normalized correlation for mouth template matching. Changes happening in the mouth region are directly described by the correlation value, i.e., the appearance of the tongue in the surface of the mouth will cause a decrease in the correlation coefficient through time. These coefficients are used for detecting the tongue protrusion. The right and left tongue protrusion positions will be detected by analyzing similarity changes between the right and left half-mouth templates and the currently tracked ones. Detection rates under the default parameters of our system are 90.20% for the tongue protrusion regardless of the position, and 84.78% for the right and left tongue protrusion positions. Our results demonstrate the feasibility of real-time tongue protrusion detection in vision-based systems and motivates further investigating the usage of this new modality in human-computer communication.

An algebraic operating system for pixel chains and codes difined for binary figures is described. Several kinds of arithmetic four operations are difined for the pixel chains and codes, and properties of the arithmetic operations and binary figures are studied to clear the problems. The algebraic operating system adopted in this paper enables us to magnify and reduce binary figures. Several examples for such operations are illustrated. In appendix, an operating system satisfying the axioms of algebraic group is described.

A quasi-TEM wave characterization method is proposed for tight couplers which are constructed of coupled microstrip lines with conductor overlay. The characterization is carried out with consideration for strip conductor thickness by applying the rectangular boundary division method and a sequential optimization procedure. Estimated characteristics of a quarter-wavelength tight coupling section are given in the graphical form for the design of structural dimensions. A trial 3-dB hybrid coupler is designed with this method and its measured characteristics are compared with theoretical ones.

A method for evaluating image segmentation methods is proposed in this paper. The method is based on a perception model where the drawing act is used to represent visual mental percepts. Each segmented image is represented by a minimal set of features and the segmentation method is tested against a set of sketches that represent a subset of the original image database, using the Mahalanobis distance function. The covariance matrix is set using a collection of sketches drawn by different users. The different drawings are demonstrated to be consistent across users. This evaluation method can be used to solve the problem of parameter selection in image segmentation, as well as to show the goodness or limitations of the different segmentation algorithms. Different well-known color segmentation algorithms are analyzed with the proposed method and the nature of each one is discussed. This evaluation method is also compared with heuristic functions that serve for the same purpose, showing the importance of using users' pictorial knowledge.

The information of closed curves obtained from figures is one of important features for describing them. Concerned with such kinds of studies, the feature extraction of a single closed curve has been reported in many papers. In this paper, a method of converting a collection of open-curves into that of the closed curves is proposed by supposing a smooth slope around the region including the curves with levels. As an example, contour lines of mountainous map are processed for obtaining closed contour lines.

Usually, paintings are more appealing than photographic images. This is because paintings can incorporate styles based on the artist's subjective view of motif. This style can be distinguished by looking at elements such as motif, color, shape deformation and brush texture. In our work, we focus on the effect of "color" element and devise a method for transforming the color of an input photograph according to a reference painting. To do this, we consider basic color category concepts in the color transformation process. We assume that color transformations from one basic color category to another may cause peculiar feelings. Therefore, we restrict each color transformation within the same basic color category. For this, our algorithm first categorizes each pixel color of a photograph into one of eleven basic color categories. Next, for every pixel color of the photograph, the algorithm finds its corresponding color in the same category of a reference painting. Finally, the algorithm substitutes the pixel color with its corresponding color. In this way, we achieve large but natural color transformations of an image.

Structure runlength coding encodes reversibly digital multi-color images. This coding processes the original images efficiently along raster scan lines. Two-dimensional structure of multi-color images is used to enhance its efficiency. The use of the structure makes the coding much more efficient than conventional runlength coding schemes. Even various gradational areas are coded effectively. This paper presents the precise coding method and compares its efficiency with that of a conventional runlength coding scheme.

A nonrecursive algorithm for address generating during the Hilbert scan of image buffer is presented. The technique is based on two level look-up tables. For large images, the additional storage is less than one percent of image buffer size. The present implementation of the algorithm is only about two times slower than the raster scan. Three level lool-up tables are generated if image data array is too big to fit image buffer and it is accessed from a mass storage. Software tools were developed for look-up tables generating and for the convenient use of the algorithm from C programs.

A graphic assembler language (GAL) for dealing with pictures is described with examples of its applications to computer graphics. Instruction codes of GAL are composed of four arithmetic operations of polynomials, control and input-output operations, and a sentence is defined for taking the coordinates of figures in a computer.

This paper presents a system for recognition of on-line cursive Hangul (Korean characters) by means of DP matching of structural information. The penalty function has the following special features. In order to prevent short spurious strokes from causing large penalties, an input stroke is weighted by its length relative to other input strokes. In order to make use of pen-up and pen-down information, a penalty is incurred when 2 strokes of differing type (i.e. pen-up with pen-down) are matched. Finally, to reduce the chance of obtaining a suboptimal solution which can result from using the greedy algorithm in DP matching, we look-ahead an extra match. In a computer simulation we obtained a recognition rate of 92% for partially cursive characters and 89% for fully cursive characters. Furthermore, for both cases combined the correct character appears 98% of the time in the top 10 candidates. Thus we confirmed that the proposed algorithm is effective in recognizing cursive Hangul.

Multi-lingual handwriting means the script written with more than one language. In this paper, a hierarchical hidden Markov model network-based approach is proposed for on-line recognition of multi-lingual cursive handwritings. Basic characters of language, language network, and intermixed use of language are modeled with hierarchical relations. Since recognition corresponds to finding an optimal path in such a network, recognition candidates of each language are combined with probability without special treatment. Character labels of handwriting, language modes, and segmentation are obtained simultaneously. However, several difficulties caused by multiple language occurred during recognition. Applied heuristic methods are Markov chain for language mode transitions, pairwise discrimination for confusing pairs, and constrained routines for side effects by language related preprocessing methods. In spite of the addition of other language, recognition accuracy of each language drops negligibly on experimental results of multi-lingual with Hangul, English, and Digit case.

Using query-by-sketch we propose an application to efficiently create collages with some user interaction. Using rough color strokes that represent the target collage, images are automatically retrieved and segmented to create a seamless collage. The database is indexed using simple geometrical and color features for each region, and histograms that represent these features for each image. The image collection is then queried by means of a simple paint tool. The individual segments retrieved are added to the collage using Poisson image editing or alpha matting. The user is able to modify the default segmentations interactively, as well as the position, scale, and blending options for each object. The resulting collage can then be used as an input query to find other relevant images from the database.

In this paper a new snake model for image morphing with semiautomated delineation which depends on Hermite's interpolation theory, is presented. The snake model will be used to specify the correspondence between features in two given images. It allows a user to extract a contour that defines a facial feature such as the lips, mouth, and profile, by only specifying the endpoints of the contour around the feature which we wish to define. We assume that the user can specify the endpoints of a curve around the features that serve as the extremities of a contour. The proposed method automatically computes the image information around these endpoints which provides the boundary conditions. Then the contour is optimized by taking this information into account near its extremities. During the iterative optimization process, the image forces are turned on progressively from the contour extremities toward the center to define the exact position of the feature. The proposed algorithm helps the user to easily define the exact position of a feature. It may also reduce the time required to establish the features of an image.

A fundamental method of producing scan animations by a digital computer is described by taking motion trajectories and lengths of light trails of objects as parameters, and examples of scan animation are illustrated.

Electrostatic drop forming characteristics of using a metal nozzle are compared with that of using glass nozzles. Taking the surface tension values of liquids as a parameter, we made experiments and concluded that the electroconductivity of liquids has prominent effects on drop formation for low surface tension liquids.

A recognition method of printed Thai characters by local information of contours is described. In the method, Freeman chain code and directional differences of contour tracing of characters are utilized for eliminating contour noises and extracting concavities and convexities of characters. Each arc is then segmented by points at which the arc bends. Several local features of arcs are extracted, and are used to detect a pair of the most similar arcs between a model and an input character. Finally, the similarity between each arc pair and the similarity between characters are calculated. Further, a model making, in which a single model is generated for each category by making use of feature values of characters belonging to the same category, is described. By applying the method to 69 categories (345 data) of 5050 dots of Thai characters rotated with five kinds of arbitrary angles, a very high recognition rate has been obtained.

In pattern recognition using neural networks, it is very difficult for researchers or users to design optimal neural network architecture for a specific task. It is possible for any kinds of neural network architectures to obtain a certain measure of recognition ratio. It is, however, difficult to get an optimal neural network architecture for a specific task analytically in the recognition ratio and effectiveness of training. In this paper, an evolutional method of training and designing feedforward neural networks is proposed. In the proposed method, a neural network is defined as one individual and neural networks whose architectures are same as one species. These networks are evaluated by normalized M. S. E. (Mean Square Error) which presents a performance of a network for training patterns. Then, their architectures evolve according to an evolution rule proposed here. Architectures of neural networks, in other words, species, are evaluated by another measurement of criteria compared with the criteria of individuals. The criteria assess the most superior individual in the species and the speed of evolution of the species. The species are increased or decreased in population size according to the criteria. The evolution rule generates a little bit different architectures of neural network from superior species. The proposed method, therefore, can generate variety of architectures of neural networks. The designing and training neural networks which performs simple 3 3 and 4 4 pixels which include vertical, horizontal and oblique lines classifications and Handwritten KATAKANA recognitions are presented. The efficiency of proposed method is also discussed.

In the field of computer animation, effective methods are required, because of a large number of complicated line drawings. In this paper, an inbetweening method of animation characters is described for a mini-computer system.

The construction of photo-realistic 3D scenes from video data is an active and competitive area of research in the fields of computer vision, image processing and computer graphics. In this paper we address our recent work in this area. Unlike most methods of 3D scene construction, we consider the generation of virtual environments from video sequence with a video-cam's forward motion. Each frame is decomposed into sub-images, which are registered correspondingly using the Levenberg-Marquardt iterative algorithm to estimate motion parameters. The registered sub-images are correspondingly pasted together to form a pseudo-3D space. By controlling the position and direction, the virtual camera can walk through this virtual space to generate novel 2D views to acquire an immersive impression. Even if the virtual camera goes deep into this virtual environment, it can still obtain a novel view while maintaining relatively high resolution.