Title :
Hausdorff distance for robust and adaptive template selection in visual target detection
Author :
Gastaldo, P. ; Zunino, R.
Author_Institution :
DIBE, Genoa Univ., Italy
fDate :
12/5/2002 12:00:00 AM
Abstract :
Template selection is often crucial for target-detection methods; the Hausdorff distance enhances template matching by implicitly encoding the notion of resemblance. In addition, Hausdorff metrics can build up a decision map to solve the template sizing and selection problems adaptively. The low computational cost of the method fits video-surveillance applications.
Keywords :
adaptive signal processing; data compression; image coding; image matching; surveillance; television applications; Hausdorff distance; Hausdorff metrics; adaptive template selection; decision map; low computational cost; resemblance encoding; robust template selection; template matching; template-sizing; video-surveillance applications; visual target detection;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20021179
