[in Japanese]

View full abstract

An efficient coding scheme at a low bit rate for color images using adaptive Vector quantization and DPCM is proposed. RGB signals are converted into C₁, C₂ and C₃ signals by Discrete Cosine Transform (DCT). C₂ and C₃ signals, which correspond approximately to color difference signals, are band-limited and subsampled.
The three converted signals are individually encoded. Every block is composed of 8 × 8 samples of each signal. One vector or scalar quantizer, predesigned using the standard distribution of prediction error signals, is adaptively selected by the table-look-up method. Then, two dimensional prediction error signals are normalized and quantized by the selected quantizer. Computer simulation indicates that this coding scheme performs well and provides a 1-2 dB improvement in SNR over the previously reported adaptive DCT coding scheme with vector quantizers.

View full abstract

The diversification of TV production methods and the appearance of new broadcasting services have made production systems much more complex. A new universal video processor “Picot” (Picture computer) has been developed. It consists of software controled standardized hardware modules. It is a parallel processor system with about 200 processors and can process video signals sampled at 14.3 MHz of 13.5 MHz in realtime.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract