Face Images Of Various Organs Located At The Application Of Computer Network Security Research

Computer network, as an open information system, possesses many potential securityloopholes and the global network infrastructure suppliers and users are paying more andmore attention to the network security technologies. At the moment, the main threat to thenetwork security comes from identity theft. Biometric identification is regarded one of topten promising high technologies in the 21st century and is a key research area in networksecurity. Biometric identification utilizes computers to recognize some humanphysiological traits (human face, fingerprint, palm print, iris, ear shape and etc.) or somebehavior traits (voice, step) and to authenticate the identity of users. As current account andpassword identity identification system is far from enough, biometric identification systemswill have important spots in future network security research.This article studies the face identification technology in biometric identificationsystem. Strictly speaking, face identification technology includes both face location andface identification. Face location is to locate the human face in the image and then usecomputer to process the image containing human face to produce a more observable andmore recognizable high quality image. Face location is a key phase in the automatichuman face identification system. Human face identification is to match the selected humanface against the existing face records in the database and retrieve the correspondinginformation for most matched face. This requires two steps: input and output. Before theidentification phase, the human features need to be fed into the database system which isknown as the input process. Well-designed algorithms are required to meet the need for fastresponse due to the large size of database. From the perspective of face location, computerscan identify some commonality of the human face based on some geometric relationsbetween the five sense organs on the face and build a human face model to differentiatehuman face and other objects. From the perspective of human face identification,computers build a unique model for each person's face. Because there are great similaritiesbetween different persons' faces, image enhancing techniques such as edge enhancement,noise filtering, sharpening and magnification are required to emphasize specific features inthe face which helps identify the face. This article uses the YCrCb color space to carry out the complexion detection. Thedetection results (bi-valued image) will contain noises which are processed usingmathematical morphological expansion, corrosion computation. While preserving theimage's high-fidelity, it helps improve global and local image quality and usefulmeasurements, data and information can be extracted from the image to produce non-imagedescription or representation. Some attributes (the natural visual features such as a region'sbrightness, color, outline and etc.) are searched and used as the flagship attributes. Then theimage is divided into separate non-intersecting regions that have distinct features whichallow users to identify their interested organs.After the previous processing, the image is converted into a bi-valued image that isbased on the complexion color. In that image, the majority white parts form a"face"andthe detection algorithm is defined as follows:The scanning position starts at 0 row and 0 column. Scan every pixel in this row and ifthe row contains any white pixel then set the element that is corresponding to this pixel inthe horizontal coordinate flag array's to true. Set the element that is corresponding to thispixel in the vertical coordinate flag array's to true. This point is the top vertical coordinateof the face, and then go to the next step. Setting false if there have not any white pixels, andscans the next row, then repeats this step . If all points in this row are black, then mark therow above this one as the bottom vertical coordinates of the face and finish the scan.Otherwise, if the point is white, then set the element that is corresponding to this pixel inthe horizontal coordinate flag array's to true. If the point is black, skip it and scan next point.Scan the horizontal coordinate array. The first element with true value marks the leftmostcoordinate for the face and the last element with true value marks the rightmost coordinateof the face. According to the above algorithms, the rectangular region for the face can beobtained.The eyes are the most important organ in human five sense organs locating on the faceand they are also the most distinguished organ that differs substantially from the whole face.Therefore, the detection for two eyes should be carried out first. If the human face imageunder processing has an angle of tilting, then the detection of two eyes can correct the tiltby using a straight line between two eyes to determine the degree of tilt."Eye"detection algorithm: Since eyes are symmetric, only one eye detection is neededand the other eye can be located based on the symmetry. Since eyes must be on the top partof the face, the detection only needs to scan the top rightmost or top leftmost region of theface to determine the eyes position. Mouth is an important organ in the face. Mouth usually is in red color which is in contrast with complexion. Because nose has no distinct featuresthat separate it from the face, the location of mouth becomes the key to correctly locatingthe face triangle. In this system, mouth is easy to appear in the bi-valued face complexionimage. Therefore, a scheme similar to the face detection is adopted to locate the mouth.This article discusses the accurate location of human face through modeling of face,eyes and mouth. Due to the time limit we have, there are many possible future works for usto investigate. Due to the variety of facial expressions; the correct identification of humanface has always been a technical challenge. Because the organs features are extracted inadvance, multiple human faces detection and identification scheme needs some furtherresearch. There is some difficulty in real-time tracking and identification of the face. Thedetection and identification performance needs to be improved. Our system has relativelyhigh requirements on the quality of the photo template. For poor, poorly-exposed photoswithout any prior processing, it has difficulty handling them. This aspect needs somefurther improvements, too. Human face organ location in human face image is the basis foridentity identification and identity identification is the key to computer network security.Identity identification directly relates to national security, daily enterprise running andpeople's daily life who all use computer network on a daily basis. As of today, human faceidentification system is not yet widely deployed. But we have to realize that it is impactingour life and is an integral part of our life.We hope that in the future research and explorations, we can improve the ways to builda human face model and make it more accurate and more effective in order to provide amore thorough service to computer network users and promote the rapid and healthydevelopment of information security industry.