Abstract / Excerpt:

Many people nowadays migrate to other countries for their jobs and hopes of a better lifestyle. It is undeniable that traffic, street, disability signs are designed to suit the country’s local preferences, and most migrants might have difficulty understanding these signs. Most of the time, symbols alone don’t give people a clear understanding as interpretations differ from each person. Using Traffic Sign Recognition Systems, this paper aims to provide synthesized speech playback for these signs.

The research aims to develop an effective alternative for a neural network TSR system which accurately detects and identifies traffic signs, and to implement audio playback for the recognized signs. Using EmguCV, shape detection and contour extraction algorithms can be implemented in C#.

For testing and evaluation, live camera capture, as well as gathered videos and images are processed in the application. There are two types of scenes: those with traffic signs, and those without traffic signs. There should be no detection made when a scene without a traffic sign is processed in the application.

Info

Source Institution	Ateneo de Davao University
Unit	Computer Studies
Authors	Justin Dale Sandoval, Henry Villaber
Page Count	17
Place of Publication	Davao City
Original Publication Date	March 1, 2015
Tags	Contour Analysis, EmguCv, Shape Detection, Traffic Sign Recognition

Preview