Why do linear SVMs trained on HOG features perform so well?
Hilton BristowSimon LuceyPublished in: CoRR (2014)
Keyphrases
- hog features
- linear svm
- human detection
- object detection
- logistic regression
- histograms of oriented gradients
- text categorization
- support vector
- feature selection
- histogram of oriented gradients
- support vectors
- hyperplane
- svm classifier
- training process
- knn
- action recognition
- local binary pattern
- surveillance system
- pedestrian detection
- linear model
- support vector machine
- data sets
- principal component analysis
- quadratic programming
- training set