作者: F. Davoine , M. Antonini , J.-M. Chassery , M. Barlaud
DOI: 10.1109/83.480769
关键词:
摘要: Presents a new scheme for fractal image compression based on adaptive Delaunay triangulation. Such partition is computed an initial set of points obtained with split and merge algorithm in grey level dependent way. The triangulation thus fully flexible returns limited number blocks allowing good ratios. Moreover, second original approach the integration classification step modified version Lloyd (vector quantization) order to reduce encoding complexity. vector quantization implemented pixel histograms directly generated from aim comparisons between two sets involved by keeping only best representative triangles domain set. Quality coding results are achieved at rates 0.25-0.5 b/pixel depending nature retained.