Abstract
Matched filtering techniques have the important property that they can detect an object within a scene containing many other objects. However, when the target class contains statistical variations, matched filtering does not perform well. On the other hand, feature extractors for statistical pattern recognition usually operate properly only when the scene contains one object at a time, thus necessitating segmentation if the scene contains many objects.
© 1986 Optical Society of America
PDF ArticleMore Like This
Zhi-Qiang Liu and Terry M. Caelli
FQ5 OSA Annual Meeting (FIO) 1986
David A. Zweig and C. J. Morgan
TuD3 Quantum-Limited Imaging and Image Processing (QLIP) 1986
Terry Caelli and Shyamala Nagendran
THD2 OSA Annual Meeting (FIO) 1986