Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device

Texture Classification Using SVM and the Fast Algorithm of Two-dimensional Wavelet Transform

Author QinDanDan
Tutor MaZuoLiang
School Jilin University
Course Computational Mathematics
Keywords Texture Classification Two - dimensional wavelet transform Mallat algorithm Feature weighting SVM classifier
CLC TP391.41
Type Master's thesis
Year 2011
Downloads 66
Quotes 0
Download Dissertation

Texture image classification has important applications in the field of biomedical engineering , remote sensing, telemetry , industrial product testing . Than three decades , the texture image classification has been popular concern about the proposed texture feature extraction and classification methods , but the most classification algorithm widespread defects ( computational complexity , correct classification rate and other issues ) , in a extent, restricted the application of these algorithms and how to extract texture features and accurate classification remains to be further research and exploration , some new theories , new technologies continue to propose further improve the texture classification algorithm provides a broad prospect . Kingsbury Tree Complex wavelet transform domain is a hot research direction in recent years, emerging , complex wavelet transform six directions , the advantage is with approximate translation invariance and rotation invariance , has similar characteristics Gabor transform and has less redundancy , which is a good way of texture feature extraction . This paper attempts to propose a texture classification method , the choice of the two-dimensional wavelet transform fast algorithm for feature extraction , wavelet transform various the band output the l1 norm as the characteristics of the texture classification and weighting them according to the degree of dispersion of the feature itself , and then re- classification using support vector machine . As can be seen from the experiment the proposed method , the complex wavelet method the basic similar classification , but the classification time is significantly less than the complex wavelet method .

Related Dissertations
More Dissertations