원문정보
초록
영어
There are several types of transport on the road; vehicles, bikes, motorcycles (two and four wheels), kids’ bikes, construction vehicles, etc. In this paper, we present an enhanced algorithm for the detection rate to classify the two-wheelers with riding people using the normalized cross correlation (NCC) and its application of histogram of oriented gradients (HOG). HOG descriptors which are one of the well known methods for object detection are the feature descriptor using edge intensity in a local region. Using the NCC algorithm which is used to make a match from template images, local weighting values are calculated from template cell features. And the combined cell features are classified by using a boosting algorithm (Adaboost). The improvement detection rates are confirmed through experiments using bicycles and motorcycles data set for the front direction.
목차
1. Introduction
2. Histogram of Oriented Gradients
3. Normalized Cross Correlation (NCC)
4. Adaboost Sequence
5. Experimental Results
6. Conclusion
References