Image coding using sub-sampling, classification of pixel blocks and association of mapping coefficients

摘要

In order to produce a decoded image substantially identical to an original image by processing a small amount of data, utilizing a thinning circuit (subsampler) a level of compression data obtained by extracting 9 pixels of an original image by 1 pixel is limited by a level limiting circuit by extracting 2 bits thereof from an MSB. In a local decoder, the original image is predicted based on correction data to output a predicted values thereof. Then, in an error calculator, a prediction error of the predicted values with respect to the original image is calculated. The compression data is corrected by a correcting circuit in such a manner that this prediction error becomes small.