Home > System > Face Recognition System
 
Face Recognition System
 
 
 
	Face recognitiontechnology is a biometric identification technology based on human 
face feature information. With the acquisition of. With the acquisition of the camera or camera 
containingA face image or video stream, and automatically in image detection and face tracking, 
face and detected by a series of characteristics of face processingRelated technology, is often
called the face recognition, face recognition.

 

 

 

 Other biological features of face and body (fingerprint, iris etc.) as a birthright, it is unique and is not easy to copy good characteristics provides the necessary precondition for the identification, compared with other types of biological recognition, face recognition has the following characteristics:

1 non mandatory: you don't need to cooperate with image acquisition equipment, almost in the unconscious state can face images are acquired, not "mandatory sampling way";

2 non contact: users and devices do not need direct contact can obtain face image;

3 Concurrency: sorting, judging and identification in practical application scenarios for more than one person can face;

4 in addition, also accord with the visual characteristics: "knowledge" in appearance, and has the advantages of simple operation, intuitive result, concealment.

____________________________________________________________________________________________

 
The principle of face recognition technology
 
   
 
Face recognition system mainly consists of four parts, respectively is: face image acquisition and detection, face image preprocessing, face image feature extraction and matching and recognition.

Face image acquisition: different face images can pass through the camera lens collected, such as static image, dynamic image, different location, different facial expressions and so on can be a very good collection of. When the user in the acquisition of equipment of shooting range, acquisition device will automatically search and photographed face image of the user.

Face detection: face detection in practice is mainly used for the pretreatment of face recognition, namely in the image accurately calibrate the face position and size. Mode characteristics contained in face images is very rich, such as histogram feature, color feature, template feature, structure feature and Haar feature. Face detection is to pick out the useful information, and use these features to achieve face detection.

Face detection method based on the above characteristics of the mainstream Adaboost learning algorithm is adopted, the Adaboost algorithm is a classification method, the classification methods weak together, a combination of new classification method is very strong.

Face detection process using Adaboost algorithm to pick out some of the most representative face rectangle features (classifier), according to the weighted voting will be weak classifiers are constructed as a strong classifier, and then some strong classifier trained a series cascade classifier cascade structure, effectively improve the speed of detection classifier.

Face image preprocessing: for the preprocessing of face image is the result of face detection based on image processing, and ultimately serves the processes of feature extraction. System to obtain the original image because of various constraints and random disturbance, often can not directly use, must be in the early stages of image processing of grayscale correction, noise filtering, image preprocessing. For a face image, the pretreatment process including face image light compensation, gray level transformation, histogram equalization, normalization, geometric correction, filtering and sharpening etc..

Feature extraction: face recognition system can be used is usually divided into visual features, statistics of pixel characteristics, transform coefficients of face image feature, face image algebraic feature. Facial feature extraction is based on certain facial features of. Facial feature extraction, also known as face representation, it is the process of feature modeling of face. Method for face feature extraction are summed up and divided into two categories: one is the characterization method based on knowledge; another is the characterization methods of algebraic feature or based on statistical learning.

Based on the characterization methods of knowledge is mainly to obtain the characteristic data help to face classification according to the shape description of face organs and distance between them, the characteristic components usually include the Euclidean distance between feature points, the curvature and angle. The eyes, nose, mouth, face, chin and other parts, the geometrical description of the relationship between the local and they can be used as an important feature, face recognition, these features are known as geometric features. Face representation knowledge mainly includes the matching method based on geometric features and template.

The basic idea of algebraic method based on feature is the face in the airspace of high dimensional description into low dimensional frequency domain or other space description. Characterization methods of algebraic feature points based on linear projection characterization method and nonlinear projection method. Based on the linear projection methods include principal component analysis or K-L transform, independent component analysis and Fisher linear discriminant analysis method. The nonlinear feature extraction method has two important branches: feature extraction technology of nuclear and manifold learning feature extraction technique based on led.

Matching and recognition: search matching feature template storage characteristics of face image database and data in the extracted, by setting a threshold, when the similarity exceeds this threshold, the match result output. Facial feature template of face recognition is to face recognition and has been compared, be judged according to the similarity on the face identity information. This process is divided into two categories: one category is confirmed, a process image comparison, another kind is to identify, is the process of image matching contrast.

________________________________________________________________________________________

 

Application prospect

 

At present the foreign face recognition technology has been widely used in government, military, banking, social security, e-commerce, security and defense fields. With the rapid growth of control and financial trading application security entrance in China, as well as the technology mature and the social identity of the raise, face recognition will be applied in more fields:


1 enterprises, residential security and management: such as face recognition access control system, anti-theft door, face recognition

2 electronic passport and ID: Chinese electronic passport security is stepping up to plan and implement;

3 public security, judicial and criminal investigation: such as the use of face recognition system and network, a nationwide manhunt

4 bank financial: such as bank vault, safe box, ATM automated teller machines, credit cards, etc.;

5 Information Security: such as computer login, e-government and e-commerce;

6 intelligent monitoring: VIP customer recognition and location.

 

________________________________________________________________________________________

 

 

Our research project:

 

 

 

 

 

Face recognition intelligent Home Furnishing remote control system

 

 

(the more scientific, more safety, more intelligent, more high-end access control security system)

To further enhance your family life taste;

To further enhance the security of your family life;