On a fast implementation of 2D-variant of Weyl's discrepancy measure
Peter M. Roth
|Titel||On a fast implementation of 2D-variant of Weyl's discrepancy measure|
|Buchtitel||Proceedings of the 1st OAGM-ARM Joint Workshop Vision Meets Robotics|
|Note||OAGM Best Paper Award|
Applying the concept of Hermann Weyl’s discrepancy as image similarity measure leads to outstanding robustness properties for template matching. However, in comparison with standard measures this approach is computationally more involving. This paper analyzes this measure from the point of view of efficient implementation for embedded vision settings. A fast implementation is proposed based on vectorization of summed-area tables, resulting in a speed-up factor 16 compared to a standard integral image based computation.