The principle of face recognition is white:
1. Face detection du
Face detection refers to judging whether there is a face dao image in dynamic zhi scene and complex background, and separating this face. Generally, there are the following methods:
Kong zhong bo ye
① Reference template method
Firstly, design one or several standard face templates, then calculate the matching degree between the samples collected by the test and the standard templates, and judge whether there is a face through the threshold;
② face rule method
Because the face has certain structural distribution characteristics, the so-called face rule method and the method of extracting these characteristics to generate corresponding rules to judge whether the test sample contains faces;
③ Sample learning method
This method adopts the method of artificial neural network in pattern recognition, that is, the classifier is generated by learning the surface image sample set and the non-surface image sample set;
④ skin color model method
This method is based on the relatively concentrated distribution of face and skin color in color space.
⑤ Characteristic sub-surface method
In this method, all face image sets are regarded as a face subspace, and whether there are face images is judged according to the distance between the detected samples and their projections in the subspace.
It is worth mentioning that the above five methods can also be used comprehensively in the actual detection system.
2. Face tracking
Face tracking refers to the dynamic target tracking of detected faces. Specifically, the model-based method or the method based on the combination of motion and model is adopted. In addition, tracking using skin color model is also a simple and effective means.
? 3. Face contrast
Face comparison is to confirm the identity of the detected face image or search for the target in the face image database. In fact, this means comparing the sampled faces with the faces in the inventory in order to find the best matching object. Therefore, the description of face determines the specific method and performance of face recognition. There are two main description methods: feature vector and texture template:
① Eigenvector method
This method first determines the size, location, distance and other attributes of facial features such as iris, nose wing and mouth corner, and then calculates their geometric features, which constitute the feature vector describing this face.
② Pattern template method
This method stores multiple standard face image templates or face image organ templates in the database, and matches all sampled pixels with all templates in the database through normalized correlation metric during comparison. In addition, there are autocorrelation networks using pattern recognition or combining features with templates.
The core of face recognition technology is actually "local human feature analysis" and "graphic/neural recognition algorithm". This algorithm is a method of using various organs and feature parts of the face. Such as compare, judging and confirm identification parameters formed by a plurality of data corresponding to geometric relation with all original parameters in a database. Generally, the judgment time is less than 1 sec.