作者: S. Makrogiannis , G. Economou , S. Fotopoulos
DOI: 10.1016/S0031-3203(01)00130-3
关键词: Pattern recognition 、 Mathematics 、 Block Truncation Coding 、 Lossy compression 、 Lossless compression 、 Texture compression 、 Artificial intelligence 、 Data compression 、 Image compression 、 Color Cell Compression 、 Fractal transform
摘要: A new region based lossy compression scheme for color images is proposed. The segmentation method belongs to the split and merge category. Splitting carried out using watershed transform. In merging stage, a fuzzy preserving rule-based system novel one-dimensional graph structure are introduced provide accurate results with reduced computational complexity. part on Shape Adaptive DCT ΔDC correction method. quantization matrices used have been designed according properties of employed Promising perceptual low bit rate range compared previously reported methods also reported.