Multi-class Vehicle Type Recognition System
ANNPR '08 Proceedings of the 3rd IAPR workshop on Artificial Neural Networks in Pattern Recognition
Proceedings of the 20th ACM international conference on Multimedia
| Hi-index | 0.00 |
This paper proposes the application of 3 different kinds of feature extractors to recognize & classify 5 models of vehicles. These feature extractors contain of Fast Fourier transform, discrete wavelet transform & discrete curvelet transform. To justify the correct amount of each feature extractor, we perform per of the mentioned transforms to input images, precisely. The used classifier in this paper is called k nearestneighbor. The results of this test show, that the right recognition rate of vehicle's model in this recognition system, at the time of using curvelet transform (Notice, all curvelet coefficients) is 100%. For decreasing the dimension of feature vectors more & choosing the best features we've used of interclass variance criteria to intraclass variance criteria. As a result of this performance, the size of feature vectors will be extremely decreased. Then, we perform our final impact feature vectors (The best Curvelet coefficients or the best wavelet coefficients or the best Fourier coefficients) to the KNN Classifier. Also, the results of this test show, the right recognition rate of vehicle's model in this recognition system, at the time of using 0.1 of all curvelet coefficients is 100%.The comparison of the 3 proposed approaches for identifying the kind of vehicles showed that curvelet transform can extract better features among the proposed dataset.