SIFT-based matching algorithm and its application in ear recognition

Abstract

Ear recognition is an emerging biometric technology and it has great potential and broad application and development space in the field of identity verification. SIFT (Scale invariant feature transform) has the advantages of better description of the model features, maintaining the structure information, the stability of the extracted feature points, the translation scale and rotation of the image and so on. In order to improve the efficiency and accuracy of image matching, a new bidirectional matching algorithm is proposed in this paper. In the experiment, to begin with different feature points are extracted from two images. Next using the BBF-based bi-directional matching method matched all these feature points respectively. the final matches were the integrated matching correspondences. Experiments results demonstrated that the new method can improve the matching accuracy and efficiency and reduce the time consuming by 44%.