On localization of objects in the wavelet domain
Lynne L. Grewe, Richard R. Brooks · 2002
Wavelets produce sparse noise tolerant image representations. This paper explores wavelets used for object localization in target images. We propose utilizing the multi-resolution property of wavelets in an efficient object localization search. A cost metric of complexity and information measures selects the best wavelet decomposition level for object localization. Experimental results are presented.