DOI: 10.1016/J.ASOC.2015.07.048
关键词: Quantization (physics) 、 Color quantization 、 Artificial intelligence 、 Cluster analysis 、 Pattern recognition 、 Pixel 、 Linde–Buzo–Gray algorithm 、 Centroid 、 Mathematics
摘要: Graphical abstractDisplay Omitted HighlightsThe Ant-tree clustering algorithm is adapted for colour quantization.The performs the two basic operations of quantization.A threshold determines if a pixel associated to most similar cluster.Palette size can be limited, but smaller obtained an image.Computational results improve some well-known quantization methods. inspired by biological ants. This paper defines variant such perform quantization. Some features have been obtain quicker and main steps on big input set. The centroid every cluster palette once complete established, represents each original image its cluster.Computational show that error quantized images than generated other