Site Network: Personal | My Company | Vision Research Lab @ UCSB |

 

The Multiview Curve Database (MCD) has been constructed to test the performance of shape/contour descriptors in the presence of perspective distortions.

 

with Dr. C. Kenney, Dr. S. Bhagavathy and Prof. B. S. Manjunath

(2004) The Multiview Curve Database (MCD) has been constructed  to test the performance of  shape/contour descriptors in the presence of perspective distortions. The images of 40 shapes (extracted from the MPEG-7 shape database) have been taken under seven different points of view. The contour of each object has been extracted, an arbitrary rotation has been applied and 7 new contours have been generated by mirroring the original ones. The database consists of 14 contours for each of the 40 objects.

 

The Database

The images of 40 shapes (extracted from the MPEG-7 shape database) have been taken under seven different points of view:

  • Central
  • Top
  • Bottom
  • Left
  • Right
  • Top-Left
  • Right-Bottom

The following images show an example of some of the shapes seen under different points of view:

Central
Top
Bottom
Right
Left
Top Left
Bottom Right
 

In total we have collected 40 different shapes taken from 7 different points of view. The contour of each shape has been extracted, an arbitrary rotation has been applied and 7 new contours have been generated by mirroring the original ones. The database consists of 14 contours for each of the 40 shapes.

The database is stored in a Matlab file .mat containing a 40x14 cell array where each entry has the following information:

lambda Intensity level at which the curve has been extracted
name Name of the image
view Name of the view
reflection 0 = original (no reflection), 1 = reflected original curve
curve 2xN matrix containing the points that sample the curve
 

Related Publications

  1. M. Zuliani, S. Bhagavathy, B. S. Manjunath, C. S. Kenney,
    "Affine-Invariant Curve Matching"
    IEEE International Conference on Image Processing, Singapore, Oct. 2004.
    VRL ID 138: [abstract] [PDF] [BibTex]

    Abstract preview: "In this paper, we propose an affine-invariant method for describing and matching curves. This is important since affine transformations are often used to model perspective distortions. More specifical..." [more]

Click here for the ICIP 2004 presentation.