Object Detection Using Generalization and Efficiency Balanced Co-Occurrence Features
Authors:
- Haoyu RenHRHaoyu Ren
- Ze-Nian LiZLZe-Nian Li
- 1 December 2015
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 46-54
Abstract
In this paper, we propose a high-accuracy object detector based on co-occurrence features. Firstly, we introduce three kinds of local co-occurrence features constructed by the traditional Haar, LBP, and HOG respectively. Then the boosted detectors are learned, where each weak classifier corresponds to a local image region with a co-occurrence feature. In addition, we propose a Generalization and Efficiency Balanced (GEB) framework for boosting training. In the feature selection procedure, the discrimination ability, the generalization power, and the computation cost of the candidate features are all evaluated for decision. As a result, the boosted detector achieves both high accuracy and good efficiency. It also shows performance competitive with the state-of-the-art methods for pedestrian detection and general object detection tasks.Keywords
This publication has 29 references indexed in Scilit:
- Handling Occlusions with Franken-Classifiers2013
- Object Bank: An Object-Level Image Representation for High-Level Visual RecognitionInternational Journal of Computer Vision, 2013
- Fast Multiple-Part Based Object Detection Using KD-Ferns2013
- Object Detection Based on Co-occurrence GMuLBP Features2012
- Pairwise Rotation Invariant Co-occurrence Local Binary Pattern2012
- Spatial pyramid co-occurrence for image classification2011
- Mining discriminative co-occurrence patterns for visual recognition2011
- Feature Extraction Based on Co-occurrence of Adjacent Local Binary Patterns2011
- Cluster Boosted Tree Classifier for Multi-View, Multi-Pose Object Detection2007
- Object recognition from local scale-invariant features1999