Human face retrieval method based on multi-resolution LBP (local binary pattern)
暂无分享,去创建一个
The invention discloses a human face retrieval method based on a multi-resolution LBP (local binary pattern), which is characterized in that: the method comprises the following implementation steps: firstly, executing a human face detection unit 20 based on multi-classifier fusion to an inputted human face image 10; then, executing a detection unit 30 of a key part in the human face image; next, executing a human face image shape normalization processing unit 40; next, executing a key part local area sampling image unit 50; next, executing a multi-resolution LBP characteristic description unit 60; next, executing a visual vocabulary generating and index building unit 70; finally, executing a similarity measurement unit 80 based on TF-IDF (term frequency-inverse document frequency); and displaying the final retrieval result in an unit 90.
[1] Peng Liang,et al. Multi-resolution local binary patterns for image classification , 2010, 2010 International Conference on Wavelet Analysis and Pattern Recognition.