Title :
Unsupervised estimation of segmentation quality using nonnegative factorization
Author :
Sandler, Roman ; Lindenbaum, Michael
Author_Institution :
Comput. Sci. Dept., Technion, Haifa
Abstract :
We propose an unsupervised method for evaluating image segmentation. Common methods are typically based on evaluating smoothness within segments and contrast between them, and the measure they provide is not explicitly related to segmentation errors. The proposed approach differs from these methods on several important points and has several advantages over them. First, it provides a meaningful, quantitative assessment of segmentation quality, in precision/recall terms, which were applicable so far only for supervised evaluation. Second, it builds on a new image model, which characterizes the segments as a mixture of basic feature distributions. The precision/recall estimates are then obtained by a nonnegative matrix factorization (NMF) process. A third important advantage is that the estimates, which are based on intrinsic properties of the specific image being evaluated and not on a comparison to typical images (learning), are relatively robust to context factors such as image quality or the presence of texture. Experimental results demonstrate the accuracy of the precision/recall estimates in comparison to ground truth based on human judgment. Moreover, it is shown that tuning a segmentation algorithm using the unsupervised measure improves the algorithm´s quality (as measured by a supervised method).
Keywords :
image segmentation; image texture; matrix decomposition; NMF; feature distributions; image quality; image segmentation; nonnegative matrix factorization process; smoothness evaluation; unsupervised method; Active contours; Computer errors; Computer science; Humans; Image quality; Image segmentation; Image texture analysis; Layout; Partitioning algorithms; Robustness;
Conference_Titel :
Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on
Conference_Location :
Anchorage, AK
Print_ISBN :
978-1-4244-2242-5
Electronic_ISBN :
1063-6919
DOI :
10.1109/CVPR.2008.4587439