A Speed-Up Scheme Based on Multiple-Instance Pruning for Pedestrian Detection Using a Support Vector Machine.
Jaehoon YuRyusuke MiyamotoTakao OnoyePublished in: IEEE Trans. Image Process. (2013)
Keyphrases
- pedestrian detection
- multiple instance
- human body
- support vector machine
- multiple instance learning
- object detection
- learning problems
- real valued
- multi class
- detection rate
- concept learning
- maximum margin
- kernel methods
- regression problems
- semi supervised learning
- histogram intersection kernel
- benchmark datasets
- data sets
- relational learning
- svm classifier
- pose estimation
- feature vectors
- support vector
- training data
- pairwise
- feature selection