Generalized RBF feature maps for Efficient Detection

By Sreekanth Vempati, Andrea Vedaldi, Andrew Zisserman, C. V. Jawahar. 2010

95 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Vision and Pattern Recognition

Citations in the corpus

Random Feature Maps for Dot Product Kernels (2012, methodology): “Vempati et al. (2010) extended this idea to provide feature maps for RBF kernels of the form K(x,y) = exp ( − 1 2σ2χ2(x,y) ) where χ is the Chi-squared distance measure.”
Compact Bilinear Pooling (2015, methodology): “This was later extended for the intersection kernel, χ(2) kernel and the exponential χ(2) kernel [35, 25, 36].”
Orthogonal Random Features (2016, background): “2 Related Works Explicit nonlinear random feature maps have been constructed for many types of kernels, such as intersection kernels [15], generalized RBF kernels [22], skewed multiplicative histogram kernels [14], additive kernels [24], and polynomial kernels [11, 18].”
Spherical Random Features for Polynomial Kernels (2015, background): “[8] V Sreekanth, Andrea Vedaldi, Andrew Zisserman, and C Jawahar.”
Compact Random Feature Maps (2013, background): “Besides (Rahimi & Recht, 2007), there have been several approaches proposed to approximate other kernels such as group invariant (Li et al., 2010), intersection (Maji & Berg, 2009), and RBF kernels (Vempati et al., 2010).”
Pedestrian detection at 100 frames per second (2012, background): “Recent work on the linear approximation of nonlinear kernels seems a promising direction [16].”
Differentially Private Mixture of Generative Neural Networks (2017, background): “, generalized RBF kernels) [45] as well as polynomial kernels [33] among others.”
High-dimensional and large-scale anomaly detection using a linear one-class SVM with deep learning (2016, background): “SVMs are non-parametric learning models, whose complexity grows quadratically with the number of records [13].”
Nonparametric Part Transfer for Fine-Grained Recognition (2014, background): “However, these methods are costly during learning and prediction, despite the speed-up with explicit feature transformations [30].”
Deep One-Class Classification (2018, methodology): “Kernel-based methods scale at least quadratically in the number of samples (Vempati et al., 2010) unless some sort of approximation technique is used (Rahimi & Recht, 2007).”
Crosstalk Cascades for Frame-Rate Pedestrian Detection (2012, background): “Finally, research on optimizing other types of classifiers [1,2,3,17,18,19] is of considerable interest, however, such methods have difficulty matching the speed achieved by boosted classifiers.”
Linear-Time Learning on Distributions with Approximate Kernel Embeddings (2015, methodology): “We further compare to RKS with histogram JS embeddings (Vempati et al. 2010) (Hist JS); we also tried χ(2) embeddings, but their performance was quite similar.”