In conventional image classification methods, complementary advantages between various single features of images are not fully applied; meanwhile, redundant information exists in the extracted features. As a result, accuracy of image classification is not high. Therefore, a novel approach for image classification based on multi-feature combination and PCA-RBaggSVM (principal component analysis and random bagging support vector machine) is proposed in the paper. First, comprehensive features describing fully image content are extracted, then redundant information is removed by transforming extracted features with PCA. Finally, RBaggSVM of ensemble SVM classifier is applied for classification. Experimental result shows that the method has higher accuracy and faster speed of image classification than similar methods.
[1]
He Ming,et al.
PCA Based Bagging Ensemble Methods
,
2006
.
[2]
Huan Liu,et al.
Efficient Feature Selection via Analysis of Relevance and Redundancy
,
2004,
J. Mach. Learn. Res..
[3]
Markus A. Stricker,et al.
Color indexing with weak spatial constraints
,
1996,
Electronic Imaging.
[4]
Chih-Jen Lin,et al.
A comparison of methods for multiclass support vector machines
,
2002,
IEEE Trans. Neural Networks.
[5]
Guan Teng-fei.
Image classification based on transductive support vector machines
,
2007
.
[6]
Wang Xi-zhao.
Image Classification Based on Wavelet Transformation and Support Vector Machine
,
2007
.