Multi-pose face detection based on feature fusion and decision tree cascade structure

In order to improve the speed and performance of multi-pose face detection in complex patterns, a multi-pose face detection method based on feature fusion and decision tree cascade structure was proposed. The edge-orientation field features based on morphological gradient were given in the proposed method, and a AdaboostSVM algorithm based on the fusion of both Haar-like and edge-orientation field features was proposed. With the improvement of the decision tree cascade structure, the AdaboostSVM algorithm with the feature fusion was combined with the improved decision tree cascade structure to carry out the multi-pose face detection. The experimental results show that the proposed method can improve the speed and performance of the multi-pose face detection in complex patterns obviously.