FPGA-GPU architecture for kernel SVM pedestrian detection.
Sebastian BauerSebastian KöhlerKonrad DollUlrich BrunsmannPublished in: CVPR Workshops (2010)
Keyphrases
- pedestrian detection
- histogram intersection kernel
- human body
- support vector
- kernel function
- real time
- feature vectors
- parallel architecture
- hardware architecture
- object detection
- kernel methods
- hardware implementation
- support vector machine svm
- support vector machine
- privileged information
- detection rate
- feature space
- hardware design
- feature selection
- human detection
- benchmark datasets
- fpga technology
- pedestrian tracking
- field programmable gate array
- multi class
- training set
- knn
- pipelined architecture
- advanced driver assistance systems
- histograms of oriented gradients
- xilinx virtex
- three dimensional
- machine learning
- parallel computing
- viewpoint
- object recognition
- training data