Vehicle logo recognition by weighted multi-class support vector machine ensembles based on sharpness histogram features

• Published in: IET image processing Vol. 9; no. 7; pp. 527 - 534
• Main Authors: Xiao, Jianli, Xiang, Wenshu, Liu, Yuncai
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 01.07.2015
• Subjects: Algorithms; average recognition performance; colour histogram features; Construction equipment; feature extraction; Histograms; image colour analysis; image feature matching approaches; image matching; Image processing; kernel function; Logos; object recognition; road vehicles; Sharpness; sharpness histogram features; Support vector machines; SVM; traffic engineering computing; vehicle logo recognition; Vehicles; weighted multiclass support vector machine ensembles; object recognition; support vector machines; average recognition performance; traffic engineering computing; SVM; kernel function; image matching; vehicle logo recognition; image feature matching approaches; road vehicles; feature extraction; weighted multiclass support vector machine ensembles; sharpness histogram features; image colour analysis; colour histogram features
• ISSN: 1751-9659; 1751-9667; 1751-9667
• DOI: 10.1049/iet-ipr.2014.0691

Classical methods recognise vehicle logos with image feature matching approaches. Different from these methods, this study proposes a novel algorithm to recognise the vehicle logos in real time by constructing the weighted multi-class support vector machine (SVM) ensemble model to classify the vehicle logos based on sharpness histogram features. To evaluate the performance of the proposed algorithm, extensive experiments have been performed. Experimental results indicate that the sharpness histogram features proposed by the authors has better distinguishability than colour histogram features. Moreover, they show that the proposed algorithm has the best average recognition performance, and its performance is the most robust. Conveniently, the proposed algorithm can avoid the burden of choosing the appropriate kernel function and parameters comparing with multi-class SVM model.