@inproceedings{scholars7185,
           title = {Vision based motorcycle detection using HOG features},
         journal = {IEEE 2015 International Conference on Signal and Image Processing Applications, ICSIPA 2015 - Proceedings},
       publisher = {Institute of Electrical and Electronics Engineers Inc.},
           pages = {452--456},
            note = {cited By 16; Conference of 4th IEEE International Conference on Signal and Image Processing Applications, ICSIPA 2015 ; Conference Date: 19 October 2015 Through 21 October 2015; Conference Code:119504},
             doi = {10.1109/ICSIPA.2015.7412234},
            year = {2016},
          author = {Mukhtar, A. and Tang, T. B.},
            isbn = {9781479989966},
        keywords = {Classification (of information); Computer vision; Highway accidents; Motorcycles; Pattern recognition; Pattern recognition systems; Signal detection; Support vector machines; Transportation, Color characteristics; Corner feature; Crash avoidance; Detection rates; Histogram of oriented gradients (HOG); ITS applications; Mode of transport; Motorcycle detection, Image processing},
             url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-84971633792&doi=10.1109\%2fICSIPA.2015.7412234&partnerID=40&md5=507f4de72cb8c21ee3fa5e0d5f89c5d7},
        abstract = {In this paper, we present a motorcycle detection system in static images leading to its application in crash avoidance systems. Motorcycles are common mode of transport in ASEAN countries and contribute more road crashes than any other mode of transport. In our proposed system, motorbikes are detected based on the helmet and tyre color characteristics. This method involves the fusion of shape, color and corner features to hypothesize motorcycle locations in a video frame. The hypothesized locations are then classified using a support vector machine (SVM) classifier trained on histogram of oriented gradients (HOG) features of motorcycle database. The proposed technique was successfully designed and implemented on a standard PC. It was able to detect single and multiple motorcycles in videos with 96 detection rate. {\^A}{\copyright} 2015 IEEE.}
}