Group Cost-Sensitive BoostLR With Vector Form Decorrelated Filters for Pedestrian Detection.
Chengju ZhouMeiqing WuSiew-Kei LamPublished in: IEEE Trans. Intell. Transp. Syst. (2020)
Keyphrases
- cost sensitive
- pedestrian detection
- human body
- object detection
- multi class
- cost sensitive learning
- misclassification costs
- benchmark datasets
- cost sensitive classification
- human detection
- class imbalance
- naive bayes
- active learning
- class distribution
- detection rate
- histograms of oriented gradients
- three dimensional
- face detection
- computational cost
- pairwise
- object recognition
- support vector