Support vector machines for automatic target recognition using wavelet kernel
2007
The classification problem of small target is a very significant but challenging task in the field of automatic target recognition. In this paper, an enhanced support vector machine with the wavelet kernel function was proposed. In order to concentrate on the classification, It is assumed that regions containing possible targets are provided. Then the Hu's moment invariants are chosen as the feature vectors used for classifiers. Finally, the classification is performed by a support vector classifier used Db4 wavelet kernel. Compared to the Gaussian kernel classifier, simulation results show that this method leads to a more admissible result in terms of classification accuracy and robustness.
Keywords:
- Pattern recognition
- Polynomial kernel
- Kernel method
- Structured support vector machine
- Linear classifier
- Kernel embedding of distributions
- Variable kernel density estimation
- Machine learning
- Radial basis function kernel
- Artificial intelligence
- Quadratic classifier
- Computer science
- Relevance vector machine
- String kernel
- Least squares support vector machine
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
16
References
0
Citations
NaN
KQI