Feature subset selection based on co-evolution for pedestrian detection
Performance of a classification-based pedestrian detection system (PDS) is greatly affected by the features adopted in the classifier, making an appropriate subset of features necessary. The combination of different types of features can improve detection accuracy, so it it helpful to simultaneously obtain a subset and the proportion of each type for the classifier. This is difficult to achieve because of the large number and various types of features extracted to better represent pedestrians. This paper proposes a co-evolutionary method to solve the problem. In the feature subset selection method, each sub-population mapped to one type of pedestrian feature, and then all sub-populations evolved co-operatively to obtain an optimal feature subset. A strategy was designed to adjust the sub-population size adaptively in order to improve the optimizing performance. The proposed method was tested on pedestrian detection applications and in comparison to other methods, the experimental results illustrate better performance.
- Record URL:
-
Authors:
- Cao, X B
- Xu, Y W
- Wei, C X
- Guo, Y P
- Publication Date: 2011-10
Language
- English
Media Info
- Media Type: Digital/other
- Features: Figures; References; Tables;
- Pagination: pp 867-879
-
Serial:
- Transactions of the Institute of Measurement and Control
- Volume: 33
- Issue Number: 7
- Publisher: Sage Publications, Incorporated
Subject/Index Terms
- TRT Terms: Computer vision; Detection and identification technologies; Driver support systems; Genetic algorithms; Image analysis; Pattern recognition systems; Pedestrian detectors; Pedestrian safety; Pedestrian vehicle crashes; System architecture; Visual texture recognition
- Subject Areas: Data and Information Technology; Highways; Pedestrians and Bicyclists; Safety and Human Factors; I83: Accidents and the Human Factor;
Filing Info
- Accession Number: 01357090
- Record Type: Publication
- Files: TRIS
- Created Date: Nov 16 2011 2:52PM