The Quality Assessment System For Electronic Passports

Along with the development of society, the photos of passport has changed from the past paper photos to electronic photos, but the way of reviewing photos still is artificial way as past.This paper has proposed a human-computer interaction system that can evaluate brightness, definition, light, pose and background of the photo automatically, and it can filter out the unqualified photos.At preset, the research of image brightness and definition has been ripe, and the alagorithm about image brightness and definition is also more, but there is no unified standard to determine the effective degree of brightness and definition. According to the inherent characteristics of passport photos, the paper develop a set of standards by statistical analysis of a large number of qualified and unqualified photos.By the standards we can know that the brightness or definition of image is eligibility or not. When evaluating the definition, the paper proposed a program that is combined the Fourier transform and the energy gradient function, the program will be more effective and accurate to filter the blurred images from the clear images.During the evaluation of light, pose and background, the paper used the characteristics of the symmetry, and asymmetric illumination, tilt posture and complex background effectively filtered out. When extracting the background from the image, the paper cutted the photo based on the characteristics of that the people is in the middle of id photo.Experiments show that this method can effectively determine the complexity of the background.Finally, system cutted the photo based on the size of photo and location, size of the face, after cutting system evaluated the size of photo again, if the size is qualified then the photo is qualified.After testing, the system can effectively assess the photos of electronic, and divide into qualified and unqualified.