An intelligent format for compressed representation of colour graphic images

The paper describes the intelligent format CIFc for compressed representation of color graphic images (GI). The proposed format is an extension of the effective IFc format, which is based on the correlation-extremal contour methods (CECM) of character-text information recognition. The technology for obtaining descriptions in CIFc format is based on the decomposition of the original GI with the selection of the text layer and the color layer. The text layer is represented by a black-and-white image obtained as a result of binarization of the bleached GI, and is a foreground mask. The color layer includes background and color information about the foreground mask and is compressed using wavelet transform. Compression of the text layer after its vectorization is carried out by the two-criteria KECM algorithm in the self-learning mode. Image restoration is carried out in reverse order on the basis of the restoration of each of the layers. A theoretical justification is given for the effectiveness of the representation of color GI in the proposed CIFc format in comparison with the closest analogue of DjVu. The theoretical justification of the effectiveness of the proposed CIFc format in comparison with the closest analogue of DjVu and the results of experiments on model and real color GI’s are given.