| Along with the network technology,computer technology,electronic technology is gradually mature,human means to convey information has become diversified.The main way is image and voice now,through the image to pass information are accounted for above 60%of all the methods,but through the voice to obtain information are accounted for less than 20%,so the main avenue of information source is through the image acquisition.Image processing,image recognition and image retrieval are the key research objects in the field of information and electronic,at present,many domestic and foreign scholars,scientists have been devoted to research in the field of image recognition,the state key universities,research institute are also active in research and development image recognition and retrieval system,human face image retrieval is the hotspot of the research.In order to solve the defect of complex calculation,retrieval speed is slow,weak robustness of traditional image retrieval,the paper study on the basis of the image retrieval theory,using a semantic based image retrieval method,this method contains the advantages of the face segmentation,feature description,pattern recognition,image retrieval method.Using digital image processing technology,face recognition technology and database technology,make the text-based image retrieval technology and content-based image retrieval technology integration as a whole,this can avoid the cumbersome process to matching image,shorten the time of retrieval.The basic idea is that,firstly detection the image of the face,analysis whether there is a face or not then localization face;secondly pretreatment face,facial features in the face region,establishment the geometric model of facial features,extraction the main feature information,and encoding the feature information;finally attached semantic information which contained the characteristics to the image,quickly retrieve a target image in a group of images with the same semantic encoding. |
PDF Full Text Request