This paper discribes a method to generate of coastline datasets for earth observing satellite. The proposed method reserves the reduced coastline by weighting expansion of the fine coastline and thresholding and thinning.

View full abstract

In this paper, we propose the method of cloud classification from GMS images. It is a thresh-olding model which use image features such as brightness temperatue, temperatue difference, temperatue variance and correlation. Using this model, each pixel in cloud region is classified into one of 9 cloud types.

View full abstract

GMS-5/VISSR data is much huge, 2.6GB a day. To use the data in near real time, fast geometric correction is needed. So we apply bilinear interpolation to geometric correction, and the interface which geometrically corrected GMS data has been designed.

View full abstract

A method for locating buried pipes by processing of pulse radar images is presented. The method employs LOG filter to extract edges representing weak signals originated from deeply buried pipes, and the Hough Transform of the extracted edges gives the microwave velocity in the ground.

View full abstract

A method for determining the optical geometry of a viewing camera relative to a glass tank is presented. The method locates scales viewed through air first, and then it locates the glass tank by comparing the scale image viewed through air and water with the first one.

View full abstract

3CCD camera has been developed obtained three kinds of different focused images simultaneously. Using optical unit technologies for 3CCD color camera, 3CCD can be exactly fixed on the prism, slightly separate to the optical axis. Various improvement on signal processing technologies have been performed, obtaining useful characteristics for image measurement. The multi-focused camera could be useful for high accuracy 3D depth measurement and so on.

View full abstract

[in Japanese]

View full abstract

A mid-air part of a potato was searched out as a tomogram by the computer processing using a method of successive approximation. In the result, the potato having the mid-air part showed remarkable difference compared with the normal one. The resolution is 80×80 line towards the section size, 30×30mm of the potato.

View full abstract

This paper presents a detection and recognition method of human movements from a series of image data. We propose a method to transform the appeared image size of human body into a real size in order to recognize the object position and locus of motion in a real space.

View full abstract

We propose a new debris flow detection system based on MPEG video. The system consists of a MPEG video camera and a computer. A detection algorithm of the system utilizes the fact that a number of MPEG motion vectors are produced in accordance with an occurence of debris flow. We confirm the advantage of the system by computer simulation with some real debris flow movies.

View full abstract

A new Digital Archive Prototype System has been developed. This system include High Definition Scanner as an image capturing sub-system, and DVD-R jukebox as a storage sub-system. DVD-R has many technical merit when it is used in digital archives.

View full abstract

We developed a Digital Archives System. This System is possible to record cultural and assets in the form of high definition digital image information (5000×4000,3D) and to manage them as database.

View full abstract

We have developed a digital still camera system realized by a new camera LSI and a RISC-type CPU with the DSP function. In this system, the still image is generated by the combination of software and hardware processing.

View full abstract

We developed a digital signal Processing LSI for digital still camera. We implemented CCD signal processor, CCD timing generator, NTSC encoder, Motion JPEG processor, 32 bit RISC CPU and various kinds of I/O on one silicon. Using this LSI for digital still camera, we confirmed row power, high speed and small size.

View full abstract

A newly developed single-chip MPEG-2 Codec LSI is briefly described. This LSI is able to encode/decode video signals based on MPEG-2 MP@ML standard. Other important features include JPEG encoding/decoding functions. Only two-chip SDRAMs are required as external memories to achieve these functions and clock frequency is 54MHz.

View full abstract

The motion estimation block of MPEG-2 Codec LSI is briefly described. Motion estimation process requires a huge amount of operations. Therefore, to lower power dissipation, it is needed to reduce logic gates and RAM size with minimum penalty of picture quality degradation. The developed motion estimation block uses 240 KGates.

View full abstract

We have proposed the image data transferring method suited for USB PC camera which signal processing is performed by software, and showed the validity of the method. We named this method as "Pull-Model method" that controls image data transfer actively depended on loads of host PC. By using this method, we accomplished even or higher framerate than conventional method (Push-Model method) without image compression.

View full abstract

Recently, MPEG technology is becoming popular in multimedia world because of its encoding/decoding efficiency. In such demand, we developed a compact MPEG camera for consumer users to handle MPEG video and audio as easy as camcorder. The MPEG camera encodes/decodes image with using a newly developed MPEG-1 CODEC LSI, and audio with software. We have realized the software real-time MPEG audio coding system with a RISC-type system CPU even at trick play mode such as fast forward. This paper describes the system to achieve software real-time encoding.

View full abstract

We have successfully developed a 1chip LSI for digital VCR signal Processing of Hi8/8mm camcorder. This LSI have been developed with the object of compact and light weight, high picture quality and low power. This paper provides a brief overview of this LSI and new Camera-Video system.

View full abstract

A single chip video signal processor system for multimedia TV which is possible to receive composite video and component video signals has been studied. In this paper, we suggest one of the horizontal and vertical systems for low cost, high performance and high reliability.

View full abstract

This paper describes the development of 42 inch PALC (Plasma Addressed Liquid Crystal) prototype display. Sony, Philips, and Sharp are developing PALC display with wide-viewing angle which is using ASM (Axially Symmetric aligned Microcell mode) technology. The prototype delivers 400 cd/m2 of brightness, a contrast ratio of 100 : 1 in daylight.

View full abstract

This paper describes development of NTSC down-converter using image DSP. As NTSC down-converter is needed to display and record high definition video signal with popular tools, we developed it for HDTV. This down-converter realize to improve the picture quality and synchronous characteristic.

View full abstract

Asynchronous multi-video signals have jitter by AD-converted with same clock. In order to simulate this jitter, a method which convert time deviation of video signal to amplitude deviation of it. This method is also effective to compensate the jitter.

View full abstract

A 1/4 inch VGA format CMOS image sensor has been developed. The sensor is fabricated using a 0.6μm CMOS process technology and the cell size is 5.6μm×5.6μm. The imager operates with a single 3.3V power supply and the fixed pattern noise is less than 0.1mV for 600mV saturation voltage.

View full abstract

In this report, we introduce an improved method for overlaying arbitrarily shaped video objects. A pixel-wise composition ratio is computed based on the binary shape at the boundary of the objects. This method improves the quality of the composite picture with low computation complexity.

View full abstract

Recently digital compression technique is used for broadcasing, DVD(digital video disk) and SNG(satellite news gathering) system. MPEG2 is the most popular and effective method to compress the video and audio signal and multiplex them. Toshiba has developed MPEG2 encoder and decoder. Thie paper presents overview and feature of them.

View full abstract

A study on improved motion vector decision scheme for low bitrate coding will be shown in this paper. Based on the experiments on MPEG-4 frame-based coding, proposed decision scheme shows the possibility to reduce the redundancy remains in the area which has high prediction error power.

View full abstract

This paper proposes the algorithm of the global motion detection as one of the preprocessing for MPEG2 video coding. The global motion detection is executed in the reduced images, which ate obtained by sub-sampling after low-pass-filtering. The simulation experiments show the algorithm gives the suitable motion vectors for both the whole image and the partial images.

View full abstract

Previously we present blotch detection and restoration method. In this paper, we examine the performance of our method for the image sequence extracted from real old film. We modify two part of our method to cope with the unique problem to real film sequence. We show the simulation results and prove the effectiveness of our method.

View full abstract

[in Japanese]

View full abstract

This paper presents a method for removal of eyeglasses frame in facial images. First, we make a parametric model of general shapes of the eyeglasses frames. Second, we estimate the inertia principal axis of each parts, and extract a part of the facial image which matches the model. Finally, by replacing the gray level of extracted region using the gray level of the surrounding skin region, we get an image of a face without eyeglasses. The results can avoid effect of eyeglasses for region segmentation of facial images.

View full abstract

In this paper, we deal with a method for detecting and restoring missing data in moving image sequence. The method is composed of the global motion segmentation and the fuzzy inference. Through computer simulations, we demonstrate the usefullness and the potentiality of the prestented restoration method, and we discuss the universality of, the parameter setting for the presented restoration method.

View full abstract

At TV-Asahi We have developed a realtime CG character "Gen-San" using motion-capture in coherence of the virtual studio system. It is on-aired every evening at the weather corner of the live coverage program "Super J Channel" and is attracting attention.

View full abstract

A special video effects called "virtual competition" has been developed to show the difference between athletes in difference race as if they were racing against each other. After collecting camera-work data, the system shoots pictures with a robot camera that operates according to the data program. Data sampling is made every 8msec, and the data are interpolated every 1msec. Control of the robot camera according to the interpolated data has made it possible to achieve high quality with precision of 0.04° max (pan) and 0.01° max (tilt).

View full abstract

We constructed the CG agent design tool and evaluated the agent made by it, in order to examine the method for designing the agent with the natural appearance, while features of the person are reflected.

View full abstract

Digital multimedia applications demand real-time and high throughput data retrieval from storage devices. Future trend for storage devices, LSI memories hard disks magneto-optical disks phase change optical disks and tapes, in digital multimedia services would be discussed in this paper. This paper suggests some technical approaches which fits optical disk devices and tape devices for digital multimedia services.

View full abstract

In this paper, some key technologies to realize a Super-density Optical Disk having more than 100 Gbit/in2 are discussed. Those technologies are an blue laser, a Solid Immersion Lens, Magnetically Super Resolution and Magnetic Domain Wall Expansion read-out etc..

View full abstract

Digital Camera, which is developed for the purpose of offering a convenient image data input device to PC, becomes popular and expands to consumer users. The system design of digital camera is utilizing the most advanced technology of semiconductor devices. Here describes the background technology of digital camera system, removable media and the movement of standardization.

View full abstract

We have proposed a new concept of Home Server to realize ISDB "anytime" services. A Home Server must be able to provide two contrasting functions at the same time - - one for the processing of programs that are constantly updated, and the other for programs to be kept in a video library. The recording medium must provide high-speed random access to the former, while possessing a large recording capacity for the latter. To satisfy these two requirements, we have developed a hierarchical Storage Management System for the Home Server.

View full abstract

VRML(Virtual Reality Modeling Language) is a three-dimensional model/scene description language primarily designed for use in the Internet environment. The goals of this paper are the following : 1) to introduce VRML briefly, 2) to report the recent development of VRML, and 3) to discuss the importance of storage technology in VRML based multiuser virtual space application.

View full abstract

[in Japanese]

View full abstract

Stereoscopic images provide a strong sensation of reality because they have depth information, that which most distinguishes them from two-dimensional images. To reproduce the sensation of reality efficiently and naturally, it is important to study the psychovisual effects of 3-D images. This paper outlines the latest findings of research in this field.

View full abstract

In holographic video sytems, recent research and development of display parts which modulate laser light spatially and temporally are presented. The parts are the acousto optical modulator and the liquid crystal panel.

View full abstract

[in Japanese]

View full abstract

The features of Head mounted Display (HMD) are wide field-of-View, wide field-of-Regard and displaying stereo images. HMDs are now used in the fields concerned with virtual reality for these features. We developed STV-E which is using monochromatic CRTs and color filters, and LH-90 which has 640x480 video format inputs.

View full abstract

The DTTB(Digital Terrestrial Television Broadcasting) has been studied, and it will be expected to realize in 2000 in Japan. In this paper, the transmission method of the Japanese DTTB system is described and its schedule is also mentioned.

View full abstract

Last year, the draft specification of interim digital terrestiral TV broadcasting system was proposed. We have developped the prototype of transmission test epuipment based on draft specfication to evaluate the transmission system and its paramers. This paper describses the outline of test equipment, technical concept, construction, blockdiagram and its functions.

View full abstract

The outline of OFDM demodulation technologies for digital terrestrial television broadcasting receiver is described. Some technical key issues of OFDM receiver are various synchronization processing; i.e. clock, carrier, symbol and frame synchronization. Furthermore, cancellation technique of phase noise caused in tuner is shown.

View full abstract

This paper describes a DTV transmitting system which might be considered implementing the digital terrestrial television broadcasting(DTTB). Especially the DTV transmitter system which prevent diminishing the service aria is proposed, and the newly developed channel combiner, which can be used in adjacent channel is also proposed.

View full abstract