Fingerprint Image Quality Assessment Study

With the development of electronic computer and information technology,the identification technology using biometric have drawn more attention.Due to the uniqueness and invariability of fingerprints,it is more practical and reliable than other identification technology,and more it is now considered the main method of the biological feature recognition.This paper's major work is the research on the methods of the estimation of the fingerprint image quality.The estimation of the fingerprint image quality is the first step in the AFIS(Automatic Fingerprint Identification System), and is one of the most important ingredients which influent the performance of the AFIS.If it is directly carry on the preprocessing(binarization,thinning etc),it will result many fake feature points and finally will make error in the matching.This paper have done some experiments on the main methods of the estimation of the fingerprint image quality,and analyzes the results, then based on these methods proposed an amelioration on these method.First give the primary judgement (primary qualified and unqualified) according to ratio of the foreground area and the ratio of directional area of the fingerprint image.Then do the examination of smudginess and dryness for the unqualified image and do the examination of the singular point ,and the offset and the orientation for the qualified image,give the message to these image.The experiment proves that the method which proposed in the paper is able to make efficient judgement to the estimation of the image,and will give the useful message to the following fingerprint preprocessing.