Research On The Face Detection And Recognition Algorithm And System Implementation On Android Platform

In the field of intelligent recognition, because of its ease of use and concealment, face recognition has been widely used in many fields, such as entrance guard system, identity data matching, etc. Face detection and recognition is an important research direction of the biological feature recognition. According to the facial features, it can detect human faces from static images or videos. Then the information of the faces features will be extracted and stored. When the target face appears again, the system can recognize the specific face by matching target with the stored data. In the fields of smart video recognition and identity authentication, the recognition software on a mobile device can be used in many applications.This paper analyzes the current situation of face recognition, and the main research directions in this field. Then the key of face recognition technologies have been studied, including image preprocessing technologies, face detection algorithms, feature extraction algorithms and target classification. At last these various functions are integrated to achieve a face recognition system. This issue begins with image pre-processing in which histogram equalization and grayscale conversion algorithm have been studied. Then the face detection algorithms based on skin color and based on the Viola Jones have been discussed. Afterward, Viola Jones method is deeply analyzed to be used in detection module. For facial features extraction, the PCA feature extraction algorithm is studied. With Viola Jones detection and PCA feature extraction, the target faces can be recognized by using KNN matching algorithm in this system.In the implementation process of face recognition, the first step is to train captured face images from which the feature data are extracted and stored in database. Then K order nearest neighbor classification method is used to classify features of target faces appeared in the detection window, and finally the results of classification belonging to feature of the database is output as the recognized faces. This software system is implemented based on Android platform by building the Android development environment on the Eclipse, using the computer vision library Open CV and combining with the Android NDK and JNI technology. The human face detection and recognition system has been tested and meets the real needs.