[in Japanese]

View full abstract

[in Japanese]

View full abstract

We have developed a new user interface that uses printed materials and image scanners. The conventional watermarking methods for printed materials are difficult to apply on small image size and to detect data rapidly. The new method encodes with two levels of small adjoining image blocks. The method applies multiple times data embedding and BCH codes for data robustness. The embedding to parts, which can expect degradation of a picture, is controlled. The information is embedded on yellow component that is not conspicuous to human's eyes. With these technologies, we realized a new user interface for printed materials with high speed, high certainty, and small size image.

View full abstract

When we gaze the object, how much the amount of visual information is obtained. When the illumination conditin and the object are changed, how the amount of visual information is changed. There is a problem how to decide the relation between these quantitatively. In this paper, the theory to calculate the amount of visual information is proposed. The various effects on the amount of visual information have been calculated by the computer sisulations. The number of perceptible brightness levels and the number of minimum perceptible areas at photopic vision are obtained from the visual properties. These values correspond to the length and kind of the sign in the information theory. First, the capacity of visual information on image is calculated. Then, the amount of visual information on image is calculated. The calculations of visual information on digital image are performed to actual digital images, and the amounts of visual information are presented quantitatively.

View full abstract

In order to make clear the relation of the coded bytes of still images and their resolutions, we measured and compared the amount of coded bytes of the original images, their reduced images and enlarged images, by adopting several reduction/enlargement techniques. The coded bytes are measured by using JPEG-LS, the newest international lossless coding standard.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Connecting to the Internet on the move is not what was thought during the origin of the Internet and an underlying wired network was assumed in the design. So modifying the scenario for the mobile environment without major change in the operational structure is not a trivial one. Moreover, mobile environment is hostile; it has low bandwidth, high rate of disconnection, high error rate etc. So how to overcome these problems and how can we engage mobile agents to solve these problems and how agents can make the tasks simple are the main issue discussed the paper. After introduction of agents and their behavior we take a look how they can support WWW access in a mobile environment and the benefits accumulated thereby.

View full abstract

[in Japanese]

View full abstract

In this Paper, We analyzed for the state of existing Image Database,Introduction of XML and SVG, Built of New Database Using XML/SVG, by Examination Verified Validity.

View full abstract

[in Japanese]

View full abstract

As the display terminal, the convenience is bad because the information that can be displayed as much as the limited. Therefore, in this research, it is built the good interface that solves these problems using the 3D virtual space. The 3D virtual space realizes using the head-mounted display (H.M.D) and several sensors that are the CCD video cameras and the tracking sensors. Also, on this interface, it is built the distributed processing that realizes low-cost system. This distributed processing system distributes the process by the unit of application.

View full abstract

[in Japanese]

View full abstract

We propose a method of improvement of accuracy of reconstructed shape. When rigid object motion of the target is acquired, we can integrate the volume intersections in time sequences. This integration is equivalent to increasing the number of cameras. The rigid object motion of the target is calculated by tracking feature points on each intersection. The landmarks on the intersection are adopted as feature points, because they are observed from anywhere and any directions. In the experiments, we reconstruct shape of a simulated model. As a result, we can obtain better accuracy in shape from integrated intersection than from an intersection at one time.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

A Web page changes in the groupware from the community wear. Community wear makes an unspecified number of people the target, and groupware supports the collaboration of the specific group. That is a thing due to the progress of the remarkable network technology in recent years. However, the amounts of information contained in the Web site increase year by year. Because of that, as for the user, it has been getting difficult to find out meaningful data. This research verifies the method to visualize the information for a user which to be necessary for, and effect on awareness.

View full abstract

[in Japanese]

View full abstract

The new observation system called 3D-ISM has been developed in recent years. This system allows filming the in-vivo full-color images. The optimal segmentation technique, however, has not been established yet. We have proposed the segmentation method getting rid of an arbitrary condition or threshold in a Region-growing method so far. In this report, we present an interactive bi-directional segmentation method applying a Linear Discriminant Analysis to the Region-growing. As a result, it succeeded in extracting the crystalline lens from continuous cross-sectional images of the human eyeball, by combining two kinds of area distinction results, and also succeeded in visualization for the 3-dimensional crystalline lens.

View full abstract

Errors in a type of sheet-feed scanner appear as deformation of lines and tingeing at edges. Errors in images including known patterns were successfully detected by checking both of deformation and tingeing. Errors in images not including unknown patterns were detected either searching for narrow colored stripes with high saturation or checking the values of L*, a* and b* in CIEL*a*b* color specification in monochrome regions.

View full abstract

The result of signpost extraction has been improved by using a newly defined color parameter names Hs, which is derived from hue and saturation. This parameter successfully rejected the effect of environmental colored lighting. It was also shown that the pre-processing was effective for objects in backlight or for objects lighted by the rising and setting sun. The extraction rate was improved from 40% to 70% by the pre-processing.

View full abstract

In recent years, in the field of individual identification, identification by biometrics attention and many researches are doing. As an individual feature, a fingerprint, the retina, etc. are mentioned and it is used for discernment based on the feature. However, for extracting these features, the body must be directly contacted to a machine, and a burden a user. Then, we consider to the feature extraction of walk sequence. In this paper, get a photograph which was taken walk sequence of a pedestrian using the video camera installed on the ceiling, and in order to verify extraction and its stability for the waveform of periodic movement of a head as an individual feature from the obtained picture, the individual was discriminated. It checked that a motion of a head can be extracted by the proposal technique, and the high rate of discernment could be mentioned in individual identification as a result of the experiment.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Face detection in unconstrained environments is a difficult task with many challenges, due to the variations in face features, poses, lighting conditions and backgrounds. This paper presents a method to detect faces from the segmented regions of skin and hair color. A face is defined as the skin color regions of a collection of skin and hair color regions that satisfies two sets of geometrical rules. Hair color is assumed to be black. By using a transductive approach for skin color segmentation and selecting pixels of low intensities adaptively among achromatic pixels for hair color segmentation, reliable segmentation of skin and hair color regions from arbitrary images taken in common office environments can be obtained. Experimental results show that the proposed method has the capability to detect multiple frontal and non-frontal faces of various poses and scales, provided proper segmentation of skin and hair color regions are available. For video input, further processing of temporal information can be used to improve the detection performance.

View full abstract

In this paper, a clustering method is proposed to perform segmentation for range image of the object that has arbitrary topology. This approach consists of two stages in order to determine desirable clusters. First, while new clusters are created on demand, the parameters of each created cluster are estimated in recursive form (Stage 1).As the algorithm processes, next, the clusters that have been created at Stage 1 compete each other, and clusters that lose the competition gradually vanish (Stage 2). Consequently, since the only desirable clusters are left, the segmentation can be performed by assigning the data to the left clusters.

View full abstract

This paper presents a setting method of domains for 3-D shape modeling by using geometric information of quadric surfaces. In fitting the parametric surface model, e.g. B-splines, to whole circumferential range data that may have arbitrary topology, how the domain over that the surface model is defined can be set is significant problem. Therefore, before the surface fitting, a pre-process is performed for segmenting the range data so that the shape that the each segmented data points imply may have simple topology. Since the segmentation method can classify the range data into several domains with similar quadric surfaces, the modeling domain in 3-D space can be set by using the geometric information of quadric surfaces. Consequently, the whole circumferential shape model can be obtained by estimating surface model over the each domain.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

BPCS-Steganography determine where the local regions that are replaceable with secret data are on binary images by using a measure called complexity and embed the data into the regions, however, BPCS embed conjugation flags that lead to whether the operation called conjugate is applied or not within fixed area on binary images. Therefore it is easy to detect the signature of the embedding. We show a method for making embedded binary patterns with a conjugation flag contained. In the method, both secret data and conjugation flags are embedded by complexity segmentation.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Quite recently, some studies on text-based animation generation have been completed.In these methods, the system puts CG(Computer Graphics) characters, CG studio and CG props by the script. When CG character acts with other content, the system let it know the detail parameters such as position, shape and the way to act. In this paper, we propose the positioning restriction metadata for CG content composition and we show the way to use these information by the virtual character and the result pictures.

View full abstract

In this paper, a multilayer videoconferencing system is presented. We proposed an idea to integrate between human image and shared workspace for training room or lecture class application. The use of multiplayer has been introduced to fulfill the sense of eye contact and support the synchronous finger pointing. A control algorithm to manage all layers of a display`s content will be described.

View full abstract

Human Acoustic and Talking Aid is at once talking aid and acoustic aid, i.e., hearing aid. We can see the vowel from the shape of the resonating part , gripped position of the operator, without hearing the sound. I introduce this voice generator without electric power supply, and mention the fundamental principle and possibility for audio-visual and tactile communication.

View full abstract