A joint wavelet representation correlator is proposed as a new architecture to combine wavelets and the joint transform correlator. It performs wavelet representation preprocessing and the correlation operation simultaneously. An intensity filter used for wavelet representation is the power spectrum of the wavelet function and can easily be synthesized and displayed. Computer simulation shows that, as compared with previous joint wavelet transform correlators, its discrimination capability is better and its performance is more stable under input noise.
© 1998 Optical Society of America
Original Manuscript: March 24, 1997
Revised Manuscript: July 21, 1997
Published: January 10, 1998
Sheng Zhong, Shutian Liu, Xueru Zhang, and Chunfei Li, "Joint wavelet representation correlator for pattern recognition," Appl. Opt. 37, 374-379 (1998)