Multiclass Road Sign Detection using Multiplicative Kernel

Abstract

We consider the problem of multiclass road sign detection using a classification function with multiplicative kernel comprised from two kernels. We show that problems of detection and within-foreground classification can be jointly solved by using one kernel to measure object-background differences and another one to account for within-class variations. The main idea behind this approach is that road signs from different foreground variations can share features that discriminate them from backgrounds. The classification function training is accomplished using SVM, thus feature sharing is obtained through support vector sharing. Training yields a family of linear detectors, where each detector corresponds to a specific foreground training sample. The redundancy among detectors is alleviated using kmedoids clustering. Finally, we report detection and classification results on a set of road sign images obtained from a camera on a moving vehicle.

Files

Full Paper as PDF

BibTeX Citation

DOI

10.20532/ccvw.2013.0011

https://doi.org/10.20532/ccvw.2013.0011

BibTeX

@InProceedings{10.20532/ccvw.2013.0011,
  author =       {Valentina Zadrija and Sini{\v s}a {\v S}egvi{\' c}},
  title =        {Multiclass Road Sign Detection using Multiplicative
                  Kernel},
  booktitle =    {Proceedings of the Croatian Compter Vision Workshop,
                  Year 1},
  pages =        {37-42},
  year =         2013,
  editor =       {Lon{\v c}ari{\' c}, Sven and {\v S}egvi{\' c},
                  Sini{\v s}a},
  address =      {Zagreb},
  month =        {September},
  organization = {Center of Excellence for Computer Vision},
  publisher =    {University of Zagreb},
  abstract =     {We consider a problem of multiclass road sign
                  detection using multiplicative kernel comprised from
                  two kernels. We show that problems of detection and
                  within-foreground classification can be jointly
                  solved by using one kernel to measure object -
                  background differences and another one to account
                  for within-class variations. The main idea behind
                  this approach is that road signs from different
                  foreground variations can share features that
                  discriminate them from backgrounds. As a model, we
                  use SVM classifier, thus feature sharing is obtained
                  through support vector sharing. Training yields a
                  family of linear detectors, where each detector
                  corresponds to a specific foreground training
                  sample. However, there may be redundancy between
                  various detectors, which is accounted for using
                  k-medoids clustering technique. Finally, we report
                  detection and classification results on a set of
                  road sign images obtained from a camera on a moving
                  vehicle.},
  doi =          {10.20532/ccvw.2013.0011},
  url =          {https://doi.org/10.20532/ccvw.2013.0011}
}