| 摘要 |
<p>PURPOSE:To elevate the accuracy of estimation and encoding efficiency, by considering a spatial distribution state of the optimum forecasting system corresponding to an encoded picture element, and estimating the optimum forecasting system in the succeeding picture element time. CONSTITUTION:As for the digital multilevel picture signal input from the terminal 11, its difference is taken from a forecasting signal selected by the selector 14, among two kinds of forecasting signals which have been generated by the forecasting device 13 in the subtracter 12, and a forecasting error is obtained. The forecasting error is quantized by the quantization device 15, and is provided to the code transducer 16 and the adder 17. The adder 17 adds the forecasting signal from the selector 14 and the forecasting error from the quantization device 15, and generates a local decoding signal. Plural forecasting signals are generated by the forecasting device 13 by using said local decoding signal. The forecasting signal is compared with the local decoding signal in the optimum forecasting system deciding circuit 18, and when its difference is small, it is stored in the memory circuit as a decision result, and the accuracy of estimation of the forecasting system is raised and the encoding efficiency is elevated by making the selector 14 select the optimum forecasting system in the following picture element time from the selective signal generating circuit 21.</p> |
