Traffic Sign Recognition (TSR) is an important application that must be incorporated in autonomous vehicles. However, machine learning methods, used normally for TSR, demand high computational resources, which is in conflict with a system that is to be incorporated into a vehicle where size, cost, power consumption and real-time response are important requirements. In this paper, we propose a TSR system based on a Reduced Kernel Extreme Learning machine (RK-ELM) which is efficiently implemented in a Graphic Processing Unit (GPU). On the one hand, the inherent simplicity of ELM-based models makes possible the recognition process to be realized in a very fast and direct way. On the other hand, the computations involved in RK-ELM can be easily implemented in a GPUs so the recognition process is clearly boosted. Experiments carried out with a commonly used dataset benchmark validate our proposal.
Reduced Kernel Extreme Learning Machine for Traffic Sign Recognition
2019-10-01
1911461 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Traffic Sign Recognition Using Kernel Extreme Learning Machines With Deep Perceptual Features
Online Contents | 2017
|Traffic Sign Recognition Using Kernel Extreme Learning Machines With Deep Perceptual Features
Online Contents | 2016
|