DocumentCode :
798650
Title :
Normalized Kemeny and Snell distance: a novel metric for quantitative evaluation of rank-order similarity of images
Author :
Luo, Jiebo ; Etz, Stephen P. ; Gray, Robert T. ; Singhal, Amit
Author_Institution :
Imaging Sci. Technol. Lab., Eastman Kodak Co., Rochester, NY, USA
Volume :
24
Issue :
8
fYear :
2002
fDate :
8/1/2002 12:00:00 AM
Firstpage :
1147
Lastpage :
1151
Abstract :
There are needs for evaluating rank order-based similarity between images. Region importance maps from image understanding algorithms or human observer studies are ordered rankings of the pixel locations. We address three problems with Kemeny and Snell´s distance (dKS), an existing measure from ordinal ranking theory, when applied to images: its high-computational cost, its bias in favor of images with sparse histograms, and its image-size dependent range of values. We present a novel computationally efficient algorithm for computing dKS between two images and we derive a normalized form dKS with no bias whose range is independent of image size. For evaluating similarity between images that can be considered as ordered rankings of pixels, dKS is subjectively superior to cross correlation.
Keywords :
computer vision; image matching; image segmentation; computational cost; computer vision; cross correlation; human observer studies; image understanding algorithms; normalized Kemeny and Snell distance; normalized form; ordinal ranking theory; pixel locations; quantitative evaluation; rank-order image similarity; region importance maps; sparse histograms; Computational efficiency; Computer vision; Costs; Histograms; Humans; Performance evaluation; Pixel;
fLanguage :
English
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
Publisher :
ieee
ISSN :
0162-8828
Type :
jour
DOI :
10.1109/TPAMI.2002.1023811
Filename :
1023811
Link To Document :
بازگشت