Hand Pose Recognition Using Local Binary Patterns and Random Forests Classifier
We propose an effective real-time hand pose recognition approach using local binary patterns and random forests classifier. Firstly, we localize the hand region from the entire image using the depth map from the depth camera. Then, we extract the feature vector from the hand image using local binary patterns. This feature vector is used to train a hand pose classifier based on random forests. In our experiments, we have constructed a large database of hand pose images and verified that the proposed LBP-based feature vector and random forests classifier is outperforms the other approaches
hand pose recognition; local binary pattern; random forests
- There are currently no refbacks.