Open AccessA Vision based Indian Traffic Sign Classification
Author(s) -
Altaf Alam,
Zainul Abdin Jaffery
Publication year - 2020
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.f8775.038620
Subject(s) - computer science , artificial intelligence , traffic sign recognition , support vector machine , pattern recognition (psychology) , outlier , visualization , traffic sign , box plot , classifier (uml) , computer vision , sign (mathematics) , mathematics , mathematical analysis , statistics
In this paper, an algorithm is proposed to classify the Indian traffic sign as mandatory cautionary and informatory class. In order to complete the task, system extracted the speed up robust features (SURF) from the Indian traffic sign data, and exploited these features to train support vector machine (SVM) algorithm. Combination of SURF features and SVM classifier makes system robust for scale variation, rotation, translation and illumination variation as well as generalization is achieved. Dimension of features have been reduced by choosing a sub set of features. Whisker and box plot visualization utilized to understand the features data. Whisker plot visualization concluded about the range, skewness, median and outliers of feature data therefore, it makes the system capable to keep good features and back out from irrelevant features. Feature refinement reduces the computational complexity. The results evaluated narrate that the overall performance of proposed algorithm is efficient.