TY - GEN
T1 - Driving posture recognition by a hierarchal classification system with multiple features
AU - Yan, Chao
AU - Zhang, Bailing
AU - Coenen, Frans
N1 - Publisher Copyright:
© 2014 IEEE.
PY - 2014/1/6
Y1 - 2014/1/6
N2 - This paper presents a novel system for vision-based driving posture recognition. The driving posture dataset was prepared by a side-mounted camera looking at a driver's left profile. After pre-processing for illumination variations, eight action classes of constitutive components of the driving activities were segmented, including normal driving, operating a cell phone, eating and smoking. A global grid-based representation for the action sequence was emphasized, which featured two consecutive steps. Step 1 generates a motion descriptive shape based on a motion frequency image(MFI), and step 2 applies the pyramid histogram of oriented gradients (PHOG) for more discriminating characterization. A three level hierarchal classification system is designed to overcome the difficulties of some overlapping classes. Four commonly applied classifiers, including k-nearest neighbor(KNN), random forest (RF), support vector machine(SVM) and multiple layer perceptron (MLP), are evaluated in each level. The overall classification accuracy is over 87.2% for the eight classes of driving actions by the proposed classification system.
AB - This paper presents a novel system for vision-based driving posture recognition. The driving posture dataset was prepared by a side-mounted camera looking at a driver's left profile. After pre-processing for illumination variations, eight action classes of constitutive components of the driving activities were segmented, including normal driving, operating a cell phone, eating and smoking. A global grid-based representation for the action sequence was emphasized, which featured two consecutive steps. Step 1 generates a motion descriptive shape based on a motion frequency image(MFI), and step 2 applies the pyramid histogram of oriented gradients (PHOG) for more discriminating characterization. A three level hierarchal classification system is designed to overcome the difficulties of some overlapping classes. Four commonly applied classifiers, including k-nearest neighbor(KNN), random forest (RF), support vector machine(SVM) and multiple layer perceptron (MLP), are evaluated in each level. The overall classification accuracy is over 87.2% for the eight classes of driving actions by the proposed classification system.
KW - driving assistance system
KW - driving posture recognition
KW - hierarchal classification
KW - motion frequency image
UR - http://www.scopus.com/inward/record.url?scp=84946531014&partnerID=8YFLogxK
U2 - 10.1109/CISP.2014.7003754
DO - 10.1109/CISP.2014.7003754
M3 - Conference Proceeding
AN - SCOPUS:84946531014
T3 - Proceedings - 2014 7th International Congress on Image and Signal Processing, CISP 2014
SP - 83
EP - 88
BT - Proceedings - 2014 7th International Congress on Image and Signal Processing, CISP 2014
A2 - Wan, Yi
A2 - Sun, Jinguang
A2 - Nan, Jingchang
A2 - Zhang, Quangui
A2 - Shao, Liangshan
A2 - Wang, Lipo
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2014 7th International Congress on Image and Signal Processing, CISP 2014
Y2 - 14 October 2014 through 16 October 2014
ER -