The carrier to noise ratio (C/N) is used for the video reception quality evaluation. It is influenced considerably by the meteorological phenomena and the noise temperature due to the rainfall. The carrier C falls down due to the rain attenuation and rain noise temperature has a great value as regards the noise N. The measurement is carried out with different antenna aperture depends on the types. Several new phenomena became known the results of the observations.
we report noise reduction of video and audio of Direct Satellite Broadcast reception, using an MSW S/N enhance and regeneration digital audio control bit. The MSW S/N enhancer which is constructed of two HSSW filters, can improve 2 to 6dB of video SN ratiounder the threshold level. Regeneration of digital audio control bit, is to exchange a range bit and a frame synchronism bit mistakenly received under 4 to 5dB of CN ratio withright bits respectively reproduced in the processor, can eliminate unpleasant noise.
This system determine the 3-D position of mobile station by GPS (Global Positioning system), and it is automatically directed FPU antenna to other side by the result. At that time, this system calculate profile between sending and receiving points, and can judge on the instant possible to relay broadcast from the area. Effect of this system was proved by field test.
A study on a coding scheme which is appropriate for motion compensation prediction (MC) error signals is presented. MC error signals have very small correlation between adjacent pixels, especially in the vertical direction. Moreover, the correlation varies to a large extent depending on scene, local feature and type of movement. Therefore, with respect to coding gain and decoded picture quality, an adaptive coding scheme where orthogonal transform or subband partition is applied only for cases having high correlation is likely efficient for MC error signals.
NHK Science and Technical Research Laboratories are studying an advanced facsimile broadcasting system using the 12 GHz band satellite television broadcasting system. In this paper, a hierarchical coding and transmission method suitable for the facsimile broadcasting system is proposed. The characteristics of this method are as follows: - Facsimile picture is composed of the standard resolution(8 dots/mm) image of A4 size and the supplenmentary images, which make up high resolution(16 dots/mm) image of the minute image segment and the half - tone image segment on the standard resolution image. - The supplementary images are coded by the conventional MH or MR coding scheme, which can be decoded very easily in the facsimile broadcasting terminal. - In comparison with the coded data amount when the same facsimile program is transmitted at a resolution of 8 dots/mm and 16 dots/mm by MR coding, the transmission efficiency is 22% better.
This paper clarifies theoretically that a 3-layerred linear perceptron with a hidden unit for identity mapping has no local minima but a global minimum in whole variable-space of network coefficients including thresholds and that after convergence the hidden unit outputs the most dominant K-L transform component of each sample among an arbitrary property of input sample-set including ones with bias, irregular covariance matrix and/or multiplicative eigenvalues of the matrix. These results suggest the feasibility of transforming input original samples to KL transform set by a multi-tandem network of Unit perceptrons.
This paper clarifies that a Multi-Tandem Perceptron (MTP) outputs the dominant K-L transform components of each sample of arbitrary input samples set. A R-stages MTP consists of R Unit-Perceptrons (UPs). Each UP is a 3-layerred linear perceptron with a hidden unit for identity mapping and transmits the mapping error vector to the following one. The hidden unit of the r-th UP outputs the r-th dominant component. This paper also discribes a rapidly learning method. The simulation results indicate that the method performs more than 100 times as rapidly as the conventional back propagation.
In this paper we describes the inevitable problems in enlarging the scale of neural networks for character recognition and approaches to them. At first we propose an initial setting rule of network coefficients, a method of avoiding inactive states in the initial learning period and a reduction algriosm of hidden units. And then we simulate neural network operations for recognizing drawn Chinese characters and alphabets by using above techniques. The simulation results indicate expected performances.
In this paper, we try to evaluate the performance and extract problems in simulating neural networks by loosely coupled multiprocessors, "Transputers" and parallel programing language, "Prallel C". At first, we propose two kinds of parallel a algorisms for neural network simulations. And then we actually simulate neural network operations by using them. Through these simulations we clarify that those algorisms realize the expected performances on the Transputers. We also indicate some remarkable problems in case of the simulations.
A HD front projector using 1.38-Mega pixels poly-Si TFT LCLVs(Liquid Crystal Light Valves) has been developed. The optical system includes dichroic prisms in the color composition system in order to reduce asymmetric aberrations. In the circuit system, motion adaptive scanning line interpolation technique is adopted for realizing the performance of Poly-Si TFT LCLVs. As a result we achieved vertical and horizontal resolution of 950 and 850 TV lines respectively.
The leakage-current characteristics of laser induced crystallized ploy-Si TFT fabricated below 450℃ have been investigated. The trap state density of laser induced crystallized poly-Si is almost the same as poly-Si fabricated around 600°. Reduction of the leakage-current was realized at higher signal voltage by the geometrical optimization. The improved characteristics permit to apply higer signal voltage for addressing pixels of normary white TN and LC/Polymer composite mode.
A TFT-LCD system suitable for multi-media applications is described. It can produce CA and AV images simulateously on a SHARP's standard CA panel that has VGA display format. A 6-bit digital source driver that can generate 64 grey levels for each color is incorporated into this panel. The system utilizes a pre-controller that compresses normal 8-bit video data into 6-bit data using the Inter-Frame Modulation Method. This method is also described.