[in Japanese]

View full abstract

The important thing in order that the CRT still shines brightly for the coming digital era might be to realize a super-lightweight and ultra thin CRT, keeping its advantages such as high image quality and low cost. One of the reasons that CRT is heavy and fragile is due to the high density of CRT glass resulted from its high X-ray absorption ability. However, it might become an impetus for realizing such an innovative CRT by introducing new CRT glass with lower density and lower brittleness. This could be possible if a new electron gun with excellent focusing characteristics even at very low accelerating voltages would be developed. In addition, we have had some advances in strengthening glass material such as field-assisted chemical tempering and improvement of the mechanical strength of a crystalline solder glass. These also will produce a possibility to innovate CRTs.

View full abstract

To reduce the cost of producing 3D computer graphics (3DCG) educational contents for e-learning systems and to improve the capability of self-learning systems, we developed a new self-learning system based on a Cyber Person, a human-shaped software robot made by using 3DCG and speech synthesis. Our system includes a 3DCG stage, many stage parts (e.g., 3DCG shaped models, photographs, text panels), and Cyber Person (s). Because the Cyber Person acts as a virtual teacher, we named it Cyber Assistant Professor (CAP). The educational contents of CAP can be written in Cyber Person Scenario Language 2 (CPSL2). CPSL2 is a tag-based hypertext language, like HTML, that describes a scenario of 3DCG interactive animation. A CAP browser converts this scenario to real-time interactive 3DCG animation, which is not a rendering movie. CAP can instruct the educational subject by using body actions, facial expressions, and speaking. The stage parts can also move simultaneously corresponding to the CAP's instructions. Sometimes, CAP shows a question to the learner and the learner has to answer by clicking the specified stage parts, e.g., selecting the correct answer or entering the correct number. In response, CAP can change the scenario of 3DCG animation or load the applicable external scenario file. We have used to develop some teaching materials for elementary school students.

View full abstract

Many TV programs broadcast captions for hearing-impaired people. Digital terrestrial television started in December 2003, so more Hi-Vision programs will be broadcast. Recently, since captions for Hi-Vision programs are produced using those for NTSC programs, the captions for Hi-Vision programs have been displayed in the same way as those of NTSC programs without modification. However, there are some differences between the viewing conditions for Hi-Vision and NTSC programs such as the screen aspect ratio and the preferable viewing distance. The presentation method of captions suitable for Hi-Vision needs to be investigated. We researched the letter size of captions and number of letters per line using a subjective evaluation test.

View full abstract

A micro-computed tomography (micro-CT) system was developed using a 10-megapixel CCD camera for 3D biomedical imaging. The system consists of a beryllium window for X-ray incidence, a fluorescent screen, an optical mirror, high numerical aperture lenses and the CCD camera. X-rays are converted into a visible image in the screen which has a thickness of 10μm. The mirror deflects the luminescent light 90° to the lens system that focuses it on the camera. The X-ray field of view is 24-mm wide by 16-mm high (4, 024×2, 648 pixels); a pixel measures 6×6μm. A rabbit auricle specimen implanted with a carcinoma was fixed in formalin after a barium sulphate solution was injected into the blood vessels as a contrast agent. 3D images of the specimen were obtained by tomographic reconstruction, and small blood vessels induced by the tumor were seen.

View full abstract

The progress of digitization at broadcasting stations enables intellectual production of high-quality content from video materials. Metadata plays an important role in describing the semantics of multimedia content, which helpssemantic structuralization of video materials. We developed an annotation method for video content based on the concept of the semantic graph, which enables semantic queries on video content. The method incorporates three ideas : metadata on video scenes, time-series of video scenes, and semantic links to external data. In the method, metadata are provided by human annotations. Human annotation is one of the practical ways based on an actual situation in TV productions. We also demonstrated its application using a real baseball game content, and evaluated the proposed annotation method.

View full abstract

In image processing, selecting image processing filters to satisfy the conversion from an original image to its target image is difficult. We previously developed an image processing system, ACTIT, that constructs image processing filters to satisfy the given conversion from original to target image. ACTIT uses GP, genetic programming, and it constructs tree-structured image processing filters with pre-defined well-known filters including several, which require numerical parameters. However, optimizing both structural and parameters at the same time is difficult for conventional GP. Thus, we developed PT-ACTIT, which can optimize structural and numerical parameters simultaneously. GA, genetic algorithm, was used to optimize nuemrical parameters. Experimental results using PT-ACTIT show that we can construct a new filter that cannot be constructed using only pre-defined filters.

View full abstract

To expand images, linear interpolation methods or cubic spline methods have generally been used. However, the expanded images generated by the methods are blurred because of a lack of high frequency components in the image signals. To obtain clear expanded images, methods based on DCT coefficients, wavelet cOefficients, or the Laplacian pyramid have been proposed. Because these methods generate or emphasize the high frequency components of image signals, problems arise such as ringings, block noise, or noise in flat parts of images. We developed a method based on a new apploach using the curved surfaces of local images expressed by polynomials and the subspace of a high-resolution image. We expect it to solve the above problems.

View full abstract

An acceleration method is presented for searching for similar images based on the lower bound of the Euclidean distance between feature vectors. The lower bound is calculated with lower dimensional feature vectors that are obtained by partitioning each image into several blocks. The effectiveness of the proposed method was determined experimentally using a simple image database. The time required for a 10-NN search was reduced by 95 percent on average.

View full abstract

Many systems to aid communication have been developed for seriously physically handicapped people, such as ALS patients. An eye-gaze input system is currently being evaluated as a new interface. This system can be used to operate a computer with eye movement only. Most conventional eye-gaze input systems use infrared rays on eyes to detect eye-movement. However, prolonged irradiation could damage eyes. We developed a new eye-gaze input system with multi-indicators using a personal computer and a home video camera to detect eye-gaze under natural light. This system does not require special devices for detecting eye-movement. The results of our system analisys show the limit of the vertical resolution is between 5 and 7 indicators.

View full abstract

An anisotropic shock filter is presented for edge-preserving enhancement of flow patterns in images. The filter consists of an anisotropic smoothing filter and a shock filter. Edges in images are sharpened in the direction of the principal eigenvector of the structure tensors. The filter is suitable for enhancing and visualizing flow patterns in images. This characteristic of the filter is used for actively generating flow patterns in images, and this technique is applied to non-photorealistic rendering for transforming photographs into oil-painting-like images or crayon-scratch-like pictures.

View full abstract

We present a variational motion de-blurring method that minimizes the regularized energy functional defined with a spatially variant model of motion blurs. Unlike spatially invariant image blurs, minimization cannot be achieved in a closed non-iterative way, and we therefore need to derive its iterative algorithm. The standard regularization method uses a square function to measure the energy of its solution function, and utilizes the energy functional composed of the data-fidelity energy term to measure deviation of the solution function from the assumed model of motion blurs and the regularization energy term to impose smoothness constraints on the solution function. However, the standard regularization method is not necessarily appropriate for motion de-blurring, because it is sensitive to model errors and errors are inevitable in motion estimation. To improve robustness against model errors, we introduced a robust estimation function into the data-fidelity energy term.

View full abstract