A complete set of the fractal functional family is proposed in order to compress image data. These functional sets are generated by means of the mid-point displacement method followed by the orthogonalization. From computer simulation results, it is found that the presently proposed new scheme is more available than the conventional DCT one from the viewpoint of the conservation of the self-affinity in the original data.

View full abstract

We propose a new method for the assessment pattern of the left ventricular wall motion from the DSA(Digital Sub-traction Angiography) image. This method uses the 3-dimensional(3-D) tensor product expansion of a 3-D array. In this paper, we comfirm the validity of the proposed method, and improve image quality by the noise elimination and make image data compression using our method.

View full abstract

This paper derives recursive linear least-squares fixed-point smoothing algorithm using covariance information by appling a Wiener-Hopf integral equation. Here the two-dimention signal, autocavariance function of signal have model of Attasi that find express as separation function by exponential function. The algorithm is obtained for the white Gaussian observation noises. The signal process is nonstationary or stationary stochastic. Autocovariance functions of the signal and the observation noise are expressed using a degenerate Kernel. The degenerate Kernel can represent covariance functions of nonstationary or stationary stochastic processes by a finite sum of nonrandom functions.

View full abstract

Scanning conversion of interlaced pictures to progressive scanning, or deinterlacing using motion compensation is discussed. Spatio-temporal frequency spectra of interlaced artifacts are investigated. It becomes clear that interline-flicker can be removed but pairing artifact cannot be by a normal motion compensation. A deinterlacing method which uses limited motion vectors in motion compensation and adaptively perform the motion-compensated interpolation and intra-field interpolation is proposed. The algorithm and the experimental converter by the proposed method and some results are described.

View full abstract

A digital audio signal transmission system using quadrature modulation of a video carrier is examined for practical use in EDTV-II. This system requires a two-dimentional waveform equalizer for reducing ghost and crosstalk. By introducing an approximation that the multiplexed signal amplitude (PCM signal) is much smaller than that of the video signal (Its level is -26dB), we have newly developed a two-dimentional waveform equalizer having an 8bit input transversal filter with simplified architecture. It can obtain a bit error rate less than 10^<-7> with the existence of up to 6dB D/U ghost.

View full abstract

This paper describes a diversity coding system for fault-tolerant networks. The proposed technique utilizes error detecting and correcting codes constructed across logically independent virtual paths in broadband networks based on ATM. This enables a nearly instantaneous data recovery from facility failure and does not need rerouting. The system configuration, a coding format and an efficient data recovery method against link failure are presented. The given analysis includes coding block delineation performance, failure detection procedure and processing delay time. The results showed that the proposed system is quite effective for easy and quick self-healing against path failure and it also maintains a high quality transmisson performance.

View full abstract

This paper proposes a parameter designing method of dynamic Connection Admission Control(CAC) for ATM networks. Dynamic CAC uses not only negotiated traffic descriptors by users but cell flow measurement to provide high utilization while maintaing QOS standards. The estimated distribution of the cell flow measurement is renewed by using a parameter that increases the weight of the measurement exponentially. The parameter dimensioning is a key to satisfy QOS standards and to provide high utilization. We analyze the relationship between the parameter and the estimate of cell loss probability with a heterogeneous bursty traffic model, and we show numerical examples demonstrating the effect of the parameter.

View full abstract