[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

This paper describes a high-performance 3-dimensional digital color decorder, that converts composit signals into component signals conforming to the CCIR Rec. 601. In order to achieve a high performance color decorder three kinds of filter a temporal-spacial comb, a two dimensional comb, and a pass band in the frequence domain are prepared.
The outputs of these filters are adaptively adjusted to the quantities of cross-talk between the luminance and chrominance of each pixel, measured by a temporal-spacial comb and 2-dimensional comb filters, then added.
The result and output is the high picture quality chrominance signal separated from the composite signal. Component signals are then easily obtaind using the multiplication and subtraction circuits.

View full abstract

Total reflection and light-path separation are the primary methods used to detect fingerprint image. Skin condition and illumination intensity influence the image quality. We investigate the influence of these factors on fingerprint images detected by both methods. With dry skin, the reflection method cannot resolve features necessary for identification, but the light-path separation method can. With wet skin, both methods are effective. For both methods, the optimal illumination intensity depends on the skin condition. The light-path separation method is superior to total reflection method.

View full abstract

The ferrite absorbing panel is used for suppressing TV ghost images caused by wave reflection from high-rise buildings. A new practical design featuring a three-layer panel with narrow slits between ferrite tiles is proposed for the surface material of high-rise buildings. An excellent ferrite absorbing panel is obtained if proper ferrite thickness and slit dimensions are selected. A beautiful, strong, and effective three-layer absorbing panel using ferrite could be realized with a thicker surface material.

View full abstract

Since CLEARVISION broadcasting was introduced in the autumn of 1989, the Ghost Cancel Reference (GCR) signal has been multiplexed on TV signals as a part of this broadcasting. Many kinds of algorithm using the GCR signal to determine the tap gain coefficients of transversal filters can be used for ghost reduction : the Zero Forcing (ZF) algorithm, Mean Square Error (MSE) algorithm, etc. We describe here the Adaptive ZF algorithm and the results of experiments on ghost reduction using this algorithm. This algorithm adaptively controls an adjustment factor in the ZF algorithm according to the criterion function and other parameters calculated from the captured data. The ZF algorithm is simple and efficient but lacks stability and noise resistance. The Adaptive ZF algorithm is an uncomplicated way to obtain speed and stabilize its operation.

View full abstract

A videosignal-processing LSI-Picot that processes standard TV signals at video-rate has been developed. Video signal processing for broadcast demands real-time processing, which means that the processor inputs and outputs data continuously. For special video effects it also demands high sampling frequency and highly precise geometrical transformation, wave generation, etc. The Picot is a processor versatile enough to meet these demands. It performs all operations in a 7Ons cycle with a precision of at least 16 bits (partly 22 bits). This paper discusses the configuration and word-length of the programmable circuits for geometrical transformation, effect wave generation, etc. The arithmetic operation circuits and the application of the Picot are also described.

View full abstract

When satellite broadcasting programs are received on mobile in built-up urban areas, visibility of the broadcasting satellite is only about 65%. In order to obtain a high receiving rate in this situation, three diversity methods are available : space diversity of the receiving side, time diversity, and space diversity of the sending side (orbital diversity). This paper compares the ability of these methods to improve receiving rate. Time diversity and orbital diversity are superior methods; either can improve reception rate to over 90% in urban areas. It is consequently possible to provide satellite programs to people on mobile in urban areas.

View full abstract

We have developed Synthevision, which makes it possible to operate the foreground camera in image synthesizing by using the background image digitally processed according to foreground camera operation. This technique requires a defocus effect to give an out-of-focus effect images. It is also necessary : 1) to get a natural defocus effect, resembling optical out-of-focus ; 2) to change the degree of defocus smoothly ; and 3) to get a big enough defocus effect. To meet these requirements using little hardware, we developed defocus equipment using a Finite Impulse Response (FIR) digital filter with a recursive network.
This paper describes the method and evaluates the quality of its defocus effect. We have tested the equipment in program production using Synthevision, and we have confirmed the advantages of this defocus method.

View full abstract

Pattern-matching coding is one of the most promising techniques for compression of picture signals, and spatial vector quantization can be considered a kind of pattern-matching coding. In spatial vector quantization, the pattern-set is generated from training images. Excess quantization error will thus result if the statistical property of a compressed image is different from that of the training images, a condition referred to as pattern-set mismatch. Another type of pattern-matching coding, 1-dimensional noiseless universal coding, was therefore developed. With this type of coding, the pattern-set is gradually built up during the coding process and thus pattern-set mismatch does not occur. Extending this concept to 2-dimensional irreversible coding, we have developed a novel pattern-matching coding technique for the compression of still images. We have also demonstrated that this technique can compress still images more efficiently than adaptive discrete cosine transform coding.

View full abstract

In the NTSC system, high frequency characteristics of chrominance signals are degraded by simple decoding of chrominance signals transmitted by VSB (Vestigial Side Band). Color sharpness therefore deteriorates. A system is proposed that expands the bandwidth of the Q-signal (narrow bandwidth chrominance signal of the NTSC system) up to 1.5 MHz-the same as the I-signal-by field-sequential transmission of high frequency chrominance components. The design of a decoder with VSB compensation, and its improvement of picture quality, are discussed.
The characteristic required of a decoder filter are analyzed, and a suitable circuit requiring minimal hardware is proposed. As evaluated by subjective tests, field-sequential transmission and VSB compensation improved picture quality.

View full abstract

[in Japanese]

View full abstract