This Article 
 Bibliographic References 
 Add to: 
Efficient Classification for Additive Kernel SVMs
Jan. 2013 (vol. 35 no. 1)
pp. 66-77
S. Maji, Toyota Technol. Inst. at Chicago, Chicago, IL, USA
A. C. Berg, Comput. Sci. Dept., Stony Brook Univ., Stony Brook, NY, USA
J. Malik, Univ. of California at Berkeley, Berkeley, CA, USA
We show that a class of nonlinear kernel SVMs admits approximate classifiers with runtime and memory complexity that is independent of the number of support vectors. This class of kernels, which we refer to as additive kernels, includes widely used kernels for histogram-based image comparison like intersection and chi-squared kernels. Additive kernel SVMs can offer significant improvements in accuracy over linear SVMs on a wide variety of tasks while having the same runtime, making them practical for large-scale recognition or real-time detection tasks. We present experiments on a variety of datasets, including the INRIA person, Daimler-Chrysler pedestrians, UIUC Cars, Caltech-101, MNIST, and USPS digits, to demonstrate the effectiveness of our method for efficient evaluation of SVMs with additive kernels. Since its introduction, our method has become integral to various state-of-the-art systems for PASCAL VOC object detection/image classification, ImageNet Challenge, TRECVID, etc. The techniques we propose can also be applied to settings where evaluation of weighted additive kernels is required, which include kernelized versions of PCA, LDA, regression, k-means, as well as speeding up the inner loop of SVM classifier training algorithms.
Index Terms:
support vector machines,approximation theory,computational complexity,image classification,approximate classifiers,additive kernel SVM,nonlinear kernel SVM,runtime complexity,memory complexity,histogram-based image comparison,intersection kernels,chi-squared kernels,large- scale recognition tasks,real-time detection tasks,INRIA person,Daimler- Chrysler pedestrians,UIUC Cars,Caltech-101,MNIST,USPS digits,weighted additive kernels,PCA kernels,LDA kernels,regression kernels,k-means kernels,SVM classifier training algorithms,support vector machines,Kernel,Additives,Histograms,Support vector machines,Complexity theory,Piecewise linear approximation,Training,additive kernels,Image classification,support vector machines,efficient classifiers
S. Maji, A. C. Berg, J. Malik, "Efficient Classification for Additive Kernel SVMs," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35, no. 1, pp. 66-77, Jan. 2013, doi:10.1109/TPAMI.2012.62
Usage of this product signifies your acceptance of the Terms of Use.