Contour Mapping Measure

  • The Contour Mapping (CM) measure is a measure of discrepancy between two contours.
  • It is based upon a matching between all points on the two given boundaries. The matching respects the arc-length ordering of the points.

Details are available in :

  • Vida Movahedi and James H. Elder (2010), “Design and Perceptual Validation of Performance Measures for Salient Object Segmentation”, in 7th IEEE Computer Society Workshop on Perceptual Organization in Computer Vision (POCV)
    Paper –  Supplementary Material
  • Here is a video comparing CM with few other measures mentioned in the paper when used in the shape approximation algorithm to minimize error between a shape hypothesis and a ground truth reference: optim.avi
Contour Mapping Measure The code for calculating CM- the contour mapping measure CM readme
CM-PrecisionRecall The code for calculation of order-preserving precision and recall values CMPR readmePR
Salient Object Dataset A collection of salient object boundaries based on Berkeley segmentation dataset SOD SOD/index.html
POCV10 paper Design and Perceptual validation of Performance measures for Salient Object Segmentation Paper Supplementary Material
POCV10 Presentation Presentation slides with talk notes POCV10_Vida

Please send your comments, questions or bug reports to:

Vida Movahedi at
Human and Computer Vision Lab, Center for Vision Research, York University