Research And Implementation Of Key Technologies About The Images Automatic Segmentation

With the development of society and science technology, intelligent technology and intelligent equipment also appeared and applied to a lot in each part of our life. In the image segmentation and recognition research field also appear the segmentation algorithm and the new treatment methods.That have made a very good segmentation and recognition efficiency, Especially in the segmentation and character recognition of human faces great progress has been made. Image segmentation technology widely used brought great convenience for our life, learning,work, communication and entertainment, increased the flexibility and improves the work efficiency. Brought the fun of life to people. Also let people free out from the busy repetition labor. So automatic image segmentation and recognition have been widespread concern by many researchers.At present the face detection and identification system has a lot of kinds and achieved good effects. The face identification mainly has two steps,the first is feature detection,and the second is the design of face Classifier. The feature detection is an important step for face detection and identification and also has a great impact on identification rate of human face. The different ways of feature detection have different advantages and disadvantages, some of them may be complicated and need a lot of calculation and time, some may take less time but have a lower identification rate, all in all, we need to find a method that can both satisfy timeliness and identification rate. After the face detection and identification tag we need to process images that cannot satisfy certificate, the main reason is the unqualified background, automatic segmentation and replace the background, image, segmentation is the key step in the image processing to image analysis, image segmentation has been studied for decades, but there’s no good results caused by automatic segmentation, now days the image segmentation in general can be divided into the following several kinds:characteristics of threshold segmentation, edge detection and region growing and region extraction.These techniques on the gray image processing has been mature, and for color image segmentation, it also has made great progress.There are a lot of interactive image segmentation technologies, and some good results of segmentation have been achieved, but the image now can’t been automaticly decollated.In this paper, based on the requirement of automatic processing photos, analyzed n ow face detection and recognition and image segmentation technology, puts forward the automatic face detection and recognition and automatic processing photo does not meet the specifications of the image. In the process of face detection and recognition, throug h the comparison of experimental results using the principal component analysis (PC A) based on the statistical properties of the signal of the second order to extract the main in gredients with unrelated properties. And then to design the classifier based on PCA feat ure extraction and support vector machine (SVM) method is used to design the decision tree, classification and recognition. This paper improved a support vector machine (SVM) based on Euclidean distance method to design the classifier, after the process if there is a face image then it will get a mark in the middle of the face as a result, otherwise it will point the photo is illegal, then according to the face after the tag photos for background is not in conformity with the requirements for background, based on improved grabCuts finish the automatic segmentation of foreground and background. Output a image that meet the electronic photo standards in the end.