A MATLAB SMO Implementation to Train a SVM Classifier: Application to Multi-Style License Plate Numbers Recognition

NEGRI, PABLO

Image Processing On Line

IPOL

Lugar: Cachan; Año: 2018 vol. 8 p. 51 - 70

This paper implements the Support Vector Machine (SVM) training procedure proposed by John Platt denominated Sequential Minimimal Optimization (SMO). The application of this system involves a multi-style license plate characters recognition identifying numbers from ´0´ to ´9´. In order to be robust against license plates with different character/background colors, the characters (numbers) visual information is encoded using Histograms of Oriented Gradients (HOG). A reliability measure to validate the system outputs is also proposed. Several tests are performed to evaluate the sensitivity of the algorithm to different parameters and kernel functions.