作者: Shan Gai , Jiafeng Liu , Xianglong Tang
DOI: 10.2991/IJCIS.2011.4.2.14
关键词:
摘要: A new method of banknote image retrieval is proposed by using set rotated quaternion wavelet filters (RQWF) and standard transform (QWT) jointly. The robust rotationally invariant features are extracted from QWT RQWF decomposed sub-bands image. Three different sets databases used to demonstrate the effectiveness method. experimental results show that improves recognition rate 78.79% 91.22% on (16000 images) database D1, form 74.08% 94.62% (20000 D2 76.44% 88.78% (10000 D3. can also obtain a reasonable level computational complexity.