Title :
Invariant noisy texture classification with bispectrum-based features from projections
Author_Institution :
Fac. of Eng., Kagawa Univ., Takamatsu, Japan
Abstract :
The author presents a novel calculation method of invariant features based on the bispectrum from the projections of images. The invariant feature is applied to the classification of sinusoidal patterns and natural texture images suffering from similarity transformations and noise. High classification performance is obtained under arbitrary rotation, small shift and scaling, as well as considerable additive noise
Keywords :
feature extraction; image classification; noise; spectral analysis; additive noise; bispectrum; bispectrum-based features; classification performance; image projections; invariant features; invariant noisy texture classification; natural texture images; pattern recognition; rotation; scaling; similarity transformations; sinusoidal patterns classification; small shift; Additive noise; Computational complexity; Fourier transforms; Frequency; Gaussian noise; Image processing; Interpolation; Noise robustness; Pattern recognition; Phase noise;
Conference_Titel :
Image Processing, 2001. Proceedings. 2001 International Conference on
Conference_Location :
Thessaloniki
Print_ISBN :
0-7803-6725-1
DOI :
10.1109/ICIP.2001.958566
