How to Evaluate Foreground Maps

The output of many algorithms in computer-vision is either non-binary maps or binary maps (e.g., salient object detection and object segmentation). Several measures have been suggested to evaluate the accuracy of these foreground maps. In this paper, we show that the most commonly-used measures for...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2014 IEEE Conference on Computer Vision and Pattern Recognition s. 248 - 255
Hlavní autoři: Margolin, Ran, Zelnik-Manor, Lihi, Tal, Ayellet
Médium: Konferenční příspěvek Journal Article
Jazyk:angličtina
Vydáno: IEEE 01.06.2014
Témata:
Accuracy
Adequacy
Algorithms
Area measurement
Computer vision
Conferences
Current measurement
Equations
evaluation
foreground extraction
Interpolation
meta-measure
Object detection
Pascal (programming language)
Pattern recognition
saliency
Segmentation
Vectors
Volatile organic compounds
ISSN:1063-6919, 1063-6919
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The output of many algorithms in computer-vision is either non-binary maps or binary maps (e.g., salient object detection and object segmentation). Several measures have been suggested to evaluate the accuracy of these foreground maps. In this paper, we show that the most commonly-used measures for evaluating both non-binary maps and binary maps do not always provide a reliable evaluation. This includes the Area-Under-the-Curve measure, the Average-Precision measure, the F β -measure, and the evaluation measure of the PASCAL VOC segmentation challenge. We start by identifying three causes of inaccurate evaluation. We then propose a new measure that amends these flaws. An appealing property of our measure is being an intuitive generalization of the F β -measure. Finally we propose four meta-measures to compare the adequacy of evaluation measures. We show via experiments that our novel measure is preferable.
Bibliografie:ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Conference-1
ObjectType-Feature-3
content type line 23
SourceType-Conference Papers & Proceedings-2
ISSN:1063-6919
1063-6919
DOI:10.1109/CVPR.2014.39