Local phase quantization for blur-insensitive image analysis

Article ID	Journal	Published Year	Pages	File Type
10359495	Image and Vision Computing	2012	12 Pages	PDF

Abstract

âº We present a new blur insensitive texture descriptor. âº Descriptors are constructed by quantizing local frequency data. âº The method is assessed in texture and face classification experiments. âº We show clear improvements over LBP, Gabor, VZ-MR8, VZ-joint, and BIF columns. âº Descriptors give the best results in the case of both sharp and blurred images.

Keywords

Feature extraction Texture recognition Face recognition Invariant features

Related Topics

Physical Sciences and Engineering Computer Science Computer Vision and Pattern Recognition

Preview

Local phase quantization for blur-insensitive image analysis

Authors

Esa Rahtu, Janne Heikkilä, Ville Ojansivu, Timo Ahonen,