Author(s): Oliveira, G. ; Silva, F. ; Pereira, D. ; Almeida, L. ; Artero, A. [UNESP] ; Bonora, A. ; Albuquerque, V. de
Date: 2019
Persistent ID: http://hdl.handle.net/11449/184663
Origin: Oasisbr
Subject(s): Traffic signs detection; Traffic signs recognition; Characters segmentation; OCR
Made available in DSpace on 2019-10-04T12:15:40Z (GMT). No. of bitstreams: 0 Previous issue date: 2018-12-01
Detection and recognition of texts in traffic signs has been widely studied and with the advance in image capture technology has helped to improve or to create new methods to achieve this issue. In this work, we presented a method for detection, segmentation and recognition of text-based traffic signs from images analyzing and processing techniques. The results show that the computational cost and accuracy rate considering the proposed approach are acceptable to real time applications, with an execution time under 0.5 seconds, with a hit rate of 94.38% in the plate detection, 83.42% in the character segmentation and 89.23 in the digit classification.
Univ Oeste Paulista, Unoeste, Presidente Prudente, SP, Brazil
Univ Estadual Paulista, Unesp, Presidente Prudente, SP, Brazil
Univ Fortaleza, Unifor, Programa Grad Informat, Fortaleza, Ceara, Brazil
Univ Estadual Paulista, Unesp, Presidente Prudente, SP, Brazil
