A new local affine invariant feature descriptor

作者：Zhang Jie Yu^*; Chen Qiang; Bai Xiao Jing; Sun Quan Sen; Xia De Shen

来源：Journal of Optoelectronics - Laser, 2009, 20(11): 1503-1506.

• Affine invariant
• Affine invariant feature
• Autoconvolution
• Descriptors
• Feature fusion
• Generalized canonical correlation analysis
• Maximally Stable Extremal Regions
• Multiscales
• Recognition rates

A new local affine invariant feature descriptor is proposed. First, a new feature named as multiscale autoconvolution entropy (MSAE) is constructed based on multiscale autoconvolution (MSA). Second, MSAE is proved to be affine invariant. Then MSA is combined with MSAE using generalized canonical correlation analysis (GCCA) to obtain a new feature with more information which can be seen as a new local affine invariant feature descriptor. Finally, the maximally stable extremal region (MSER) is described by the new descriptor, and then put into two recognition experiments. The results show that the new descriptor has a higher recognition rate.

• 出版日期2009
• 单位南京理工大学