*Assistant Professor, Department of Computer Science and Engineering, Government College of Engineering, Tirunelveli, Tamilnadu, India
**Assistant Professor, Department of Computer Science and Engineering, Anna University, Chennai, Tamilnadu, India
Due to the high variance of pedestrians and fast-changing background, pedestrian detection has been a challenging task. Traditional HOG along with SVM methods have three major challenges. They include false positives rate, true positive rate, and miss rate. The Segment of Circle HOG have been proposed and implemented as a feature extraction for pedestrian detection is introduced to improve the performance metrics. The experimented results using Daimler Chrysler Pedestrian Benchmark dataset show that the proposed method outperforms the traditional HOG method.
Segment of Circle Histogram of Gradient, Support Vector Machine, Feature Extraction, Pedestrian Classification, Detection Rate
