Strong edge features for image coding
Document typeConference report
PublisherP. MARAGOS, R.W. SCHAFER KLUWER ACADEMIA PUBLISHERS
Rights accessOpen Access
A two-component model is proposed for perceptual image coding. For the first component of the model, the watershed operator is used to detect strong edge features. Then, an efficient morphological interpolation algorithm reconstructs the smooth areas of the image from the extracted edge information, also known as sketch data. The residual component, containing fine textures, is separately coded by a subband coding scheme. The morphological operators involved in the coding of the primary component perform very efficiently compared to conventional techniques like the LGO operator, used for the edge extraction, or the diffusion filters, iteratively applied for the interpolation of smooth areas in previously reported sketch-based coding schemes.
CitationCasas, J., Torres, L. Strong edge features for image coding. A: "Mathematical Morphology and its Applications to Image and Signal Processing" IEEE, 1996, p. 443-450.