Distance Measures for Image Segmentation Evaluation

Jiang, Xiaoyi; Marti, Cyril; Irniger, Christophe; Bunke, Horst

doi:10.1155/asp/2006/35909

Cited by 77 publications

(67 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead of using a single measure for quantifying the segmentation quality, some authors support the idea of taking a collection of similar measures for defining an overall performance measure [48]. In a global sense, our proposal has reached the best performance in terms of the quantitative measures for the segmentation goodness, obtaining the best results in three out of four of them.…”

Section: Discussionmentioning

confidence: 98%

Unsupervised colour image segmentation by low-level perceptual grouping

Martínez-Usó

Pla

García-Sevilla

2011

Pattern Anal Applic

View full text Add to dashboard Cite

This paper proposes a new unsupervised approach for colour image segmentation. A hierarchy of image partitions is created on the basis of a function that merges spatially connected regions according to primary perceptual criteria. Likewise, a global function that measures the goodness of each defined partition is used to choose the best low-level perceptual grouping in the hierarchy. Contributions also include a comparative study with five unsupervised colour image segmentation techniques. These techniques have been frequently used as a reference in other comparisons. The results obtained by each method have been systematically evaluated using four well-known unsupervised measures for judging the segmentation quality. Our methodology has globally shown the best performance, obtaining better results in three out of four of these segmentation quality measures. Experiments will also show that our proposal finds low-level perceptual solutions that are highly correlated with the ones provided by humans.

show abstract

Section: Discussionmentioning

confidence: 98%

Unsupervised colour image segmentation by low-level perceptual grouping

Martínez-Usó

Pla

García-Sevilla

2011

Pattern Anal Applic

View full text Add to dashboard Cite

show abstract

“…This idea was presented in [4] as symmetric partition-distance, in [14] as bipartite-graphmatching (BGM) distance, and in the context of clustering comparison, in [22] as classification error distance. It is shown in [4] that it is equivalent to the minimum number of pixels that must not be taken into account for the two partitions to be identical.…”

Section: Pixel-set Clusteringmentioning

confidence: 99%

Measures and Meta-Measures for the Supervised Evaluation of Image Segmentation

Pont-Tuset

Marqués

2013

2013 IEEE Conference on Computer Vision and Pattern Recognition

View full text Add to dashboard Cite

This paper tackles the supervised evaluation of image segmentation algorithms. First, it surveys and structures the measures used to compare the segmentation results with a ground truth database; and proposes a new measure: the precision-recall for objects and parts. To compare the goodness of these measures, it defines three quantitative meta-measures involving six state of the art segmentation methods. The meta-measures consist in assuming some plausible hypotheses about the results and assessing how well each measure reflects these hypotheses. As a conclusion, this paper proposes the precision-recall curves for boundaries and for objects-and-parts as the tool of choice for the supervised evaluation of image segmentation. We make the datasets and code of all the measures publicly available.

show abstract

“…1 and 2 (and Figs. 4 and 5) were used as reference ground truths (Jiang et al 2006) for those two patterns (Fernandez-Garcıa et al 2008). Then an EDA was performed for each pattern to determine their typical EGD from RGB cube diagonal.…”

Section: B Cloud and Sky Pattern Characterization With Exploratory Dmentioning

confidence: 99%

The Use of Euclidean Geometric Distance on RGB Color Space for the Classification of Sky and Cloud Patterns

Neto

Wangenheim

Pereira

et al. 2010

Journal of Atmospheric and Oceanic Technology

View full text Add to dashboard Cite

The current work describes the use of multidimensional Euclidean geometric distance (EGD) and Bayesian methods to characterize and classify the sky and cloud patterns present in image pixels. From specific images and using visualization tools, it was noticed that sky and cloud patterns occupy a typical locus on the redgreen-blue (RGB) color space. These two patterns were linearly distributed parallel to the RGB cube's main diagonal at distinct distances. A characterization of the cloud and sky patterns EGD was done by supervision to eliminate errors due to outlier patterns in the analysis. The exploratory data analysis of EGD for sky and cloud patterns showed a Gaussian distribution, allowing generalizations based on the central limit theorem. An intensity scale of brightness is proposed from the Euclidean geometric projection (EGP) on the RGB cube's main diagonal. An EGD-based classification method was adapted to be properly compared with existing ones found in related literature, because they restrict the examined color-space domain. Elimination of this limitation was considered a sufficient criterion for a classification system that has resource restrictions. The EGD-adapted results showed a correlation of 97.9% for clouds and 98.4% for sky when compared to established classification methods. It was also observed that EGD was able to classify cloud and sky patterns invariant to their brightness attributes and with reduced variability because of the sun zenith angle changes. In addition, it was observed that Mie scattering could be noticed and eliminated (together with the reflector's dust) as an outlier during the analysis. Although Mie scattering could be classified with additional analysis, this is left as a suggestion for future work.

show abstract

Distance Measures for Image Segmentation Evaluation

Cited by 77 publications

References 22 publications

Unsupervised colour image segmentation by low-level perceptual grouping

Unsupervised colour image segmentation by low-level perceptual grouping

Measures and Meta-Measures for the Supervised Evaluation of Image Segmentation

The Use of Euclidean Geometric Distance on RGB Color Space for the Classification of Sky and Cloud Patterns

Contact Info

Product

Resources

About