Person/vehicle classification based on deep belief networks
- 1 August 2014
- conference paper
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE) in 2014 10th International Conference on Natural Computation (ICNC)
Abstract
In this paper, we investigated the deep learning model for object classification. Robust classification networks were trained based on Deep Belief Networks (DBN) combined with several object representations included image pixel value, feature histogram by Histogram of Oriented Gradients (HOG) operator and eigen-features to distinguish four categories: pedestrian, biker, vehicle and others in the real scene. In addition, an image dataset called NUPTERC, in which the sample images collected from real surveillance video and Internet, was built to test the proposed methods. Experiments based on NUPTERC dataset demonstrated that the proposed deep learning architecture could achieve superior person vehicle classification performance under illumination changes, large pose variations and different resolution.Keywords
This publication has 9 references indexed in Scilit:
- Representation Learning: A Review and New PerspectivesIEEE Transactions on Pattern Analysis and Machine Intelligence, 2013
- Deep Machine Learning - A New Frontier in Artificial Intelligence Research [Research Frontier]IEEE Computational Intelligence Magazine, 2010
- Speeded-Up Robust Features (SURF)Computer Vision and Image Understanding, 2008
- Face Description with Local Binary Patterns: Application to Face RecognitionIEEE Transactions on Pattern Analysis and Machine Intelligence, 2006
- Reducing the Dimensionality of Data with Neural NetworksScience, 2006
- Histograms of Oriented Gradients for Human DetectionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- Distinctive Image Features from Scale-Invariant KeypointsInternational Journal of Computer Vision, 2004
- A real-time system for classification of moving objectsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2003
- Classification of Moving Targets Based on Motion and AppearancePublished by British Machine Vision Association and Society for Pattern Recognition ,2003