Image feature extraction with the optical Haar wavelet transform

Publisher Website
Google Scholar
Abstract
Optical wavelet transformation is implemented with a volume hologram in the frequency domain. The 2-D Haar wavelets of different types and with different dilation factors are encoded with a polarizer, and their Fourier spectra are recorded in a volume hologram with an angular multiplexing technique. Different local features (such as corners and edges) of an input image are extracted simultaneously by the corner and edge Haar wavelets. Experimental results and computer simulation are compared and presented. The limitations and the feasibility of the system are discussed.

This publication has 0 references indexed in Scilit: