作者: Berrin Yanikoglu , Alisher Kholmatov
DOI: 10.1155/2009/260516
关键词: Word error rate 、 Normalization (statistics) 、 Algorithm 、 Computer science 、 Pattern recognition (psychology) 、 Handwriting recognition 、 Dynamic time warping 、 Fourier transform 、 Normalization (image processing) 、 Fast Fourier transform 、 Blossom algorithm
摘要: We present a novel online signature verification system based on the Fast Fourier Transform. The advantage of using domain is ability to compactly represent an fixed number coefficients. fixed-length representation leads fast matching algorithms and essential in certain applications. challenge other hand find right preprocessing steps algorithm for this representation. report effectiveness proposed method, along with effects individual normalization steps, comprehensive tests over two public databases. also propose use pen-up duration information identifying forgeries. best results obtained SUSIG-Visual subcorpus MCYT-100 database are 6.2% 12.1% error rate skilled forgeries, respectively. fusion our state-of-the-art Dynamic Time Warping (DTW) lowers DTW by up about 25%. While current rates higher than these databases, as approach global features, possesses many advantages. Considering suggested improvements, FFT shows promise both stand-alone especially combination approaches that local features.