Image coding device, image decoding device, image coding/decoding device and image coding/decoding method

An image coding device, comprising:

• an image input part that inputs an image;
• plural pixel value predictors that predict a value of a target pixel to be coded in the image input by the image input part according to respective different prediction methods;
• a prediction error calculator that calculates an error between the value of the target pixel in the image input by the image input part and the value predicted according to the predetermined prediction method;
• a high frequency number determiner that judges whether or not each pixel value predicted by the plural pixel value predictors and the value of the target pixel are coincident and counts the frequency of the coincidences;
• a selector that judges whether or not each pixel value predicted by the plural pixel value predictors and the value of the target pixel are coincident and selectively outputs, if they are coincident, identification information for identifying the pixel value predictor, or, if they are not coincident, the error calculated by the prediction error calculator;
• an encoder that encodes the identification information or the error selectively output by the selector according to a predetermined method;
• and
• an output part that outputs a code encoded by the encoder, wherein
• the high frequency number determiner obtains high frequency number information including a frequency of the prediction hits as first processing, and
• the selector identifies and selects the pixel value predictor based on the high frequency number information as second processing.

An image coding method, comprising the steps of:

• inputting an image;
• predicting a value of a target pixel to be coded in the image according to plural different prediction methods;
• judging whether or not each value predicted by the plural pixel value predictors and the value of the target pixel are coincident;
• counting, if the coincidence continues, runs which are continuous lengths of the coincidences for every prediction method, and terminating the runs if the coincidence does not continue;
• when the runs of all pixel value predictors are terminated, if the length of the longest run, that continued till immediately before the termination of the run, is at least one, coding the run length and identification information for identifying the predictor;
• initializing count values acquired by counting the runs;
• and
• newly starting the counting runs, if the value predicted by at least one pixel value predictor and the value of the target pixel are coincident, and coding an error between the value of the target pixel and the value predicted according to the predetermined prediction method if the values are not coincident.