The initial processing of this image aims to improve the image by manipulating the image parameters in order to get better quality. This process is necessary because of noise or noise on the face image. In addition, the color conditions on the image of the face is too contrast, less sharp or blurred to cause the information submitted image is reduced.
After the initial image processing process will be produced image of the face that has a good quality. The next step is to localize the position of the face image. This process aims to determine whether in an image there is a face image or not, but it is also done detection of the exact location of the image of the face. Once the position of the face image is known, the next process is to extract the features. The feature extraction process is used to obtain face characteristic feature information. The second process is the matching process. This process aims to match whether a face image is recognized or not. The initial stages until the feature extraction equals the differentiating registration process is that the attributes obtained are not stored to the database, but the matching is done whether the face image is identical or similar to one of the face images stored in the database. So to find out the process of measuring similarity. All available at Security systems companies in kuwait you can see it.
The last stage is the conclusion. In general, the usual conclusion drawing techniques used is to use the value of a limit or threshold that will determine whether the value of similarity obtained has met the criteria of two facial images are said to be similar or not. If the value of similarity is beyond the specified threshold limit then the image of the face inputted together with one of the face image contained in the database.
The facial image recognition technology has been continuously developed for the purpose of producing a robust facial recognition system for disturbances, fast and accurate.