作者: Yossi Rubner , Jan Puzicha , Carlo Tomasi , Joachim M Buhmann
关键词: Computer vision 、 Measure (data warehouse) 、 Image texture 、 Pattern recognition 、 Image retrieval 、 Artificial intelligence 、 Mathematics 、 Texture (music) 、 Segmentation 、 Ground truth 、 Image segmentation 、 Scale (map)
摘要: This paper empirically compares nine families of image dissimilarity measures that are based on distributions color and texture features summarizing over 1000 CPU hours computational experiments. Ground truth is collected via a novel random sampling scheme for color, by an partitioning method texture. Quantitative performance evaluations given classification, retrieval, segmentation tasks, wide variety measure parameters. It demonstrated how the selection measure, large scale evaluation, substantially improves quality unsupervised images.