Reading 2: Sketch Recognition Algorithms for Comparing Complex and Unpredictable Shapes

Citation

Field, Martin, et al. "Sketch recognition algorithms for comparing complex and unpredictable shapes." IJCAI. 2011.

PDF

Summary

This paper provides further insights into the algorithms discussed in Mechanix paper discussed in Reading 1. This paper more focuses on the algorithms used in Mechanix to compare two sketches. The three metrics used for comparison are Hausdorff Distance, Modified Hausdorff Distance and a variant of Tanimoto Coefficient.

Discussion

Hausdorff Distance gives the maximum of all the minimum distances between any two points in the two sketches being compared. Modified Hausdorff Distance takes the mean of Hausdorff distances from A to B and B to A. Then a match probability is calculated. The third metric is a modified version of Tanimoto Coefficient which takes into account the matching from sketch A to B as well from B to A. An average of the above three metrics is computed and if it is above a certain threshold (0.65) the sketches are considered to be matched.