Multi-Task Deep Metric Learning with Boundary Discriminative Information for Cross-Age Face Verification
- PDF / 1,944,361 Bytes
- 14 Pages / 547.087 x 737.008 pts Page_size
- 44 Downloads / 241 Views
Multi-Task Deep Metric Learning with Boundary Discriminative Information for Cross-Age Face Verification Tongguang Ni & Xiaoqing Gu & Cong Zhang & Weibo Wang & Yiqing Fan
Received: 12 April 2019 / Accepted: 30 September 2019 # Springer Nature B.V. 2019
Abstract Image based face verification has attracted extension attention in the fields of pattern recognition and intelligent vision. With difference in age, cross-age face verification from facial images remains a challenging work because of a large number of facial variations caused by shape, skin color and wrinkles and so on. This study proposes a multi-task deep metric learning with boundary discriminative information method called MDML-BDI. It learns a distance metric by exploring discriminative information among the interclass neighborhood samples, such that the distances between intraclass samples are as small as possible and that between interclass neighborhood samples are as far as possible. MDML-BDI learns hierarchical nonlinear transformations by integrating metric learning into the framework of multi-task deep neural network, such that a common shared layer shares the common transformation by multiple tasks, and the other independent layers T. Ni : X. Gu (*) : C. Zhang School of Information Science and Engineering, Changzhou University, Changzhou 213164, China e-mail: [email protected] W. Wang School of Information Science and Engineering, East China University of Science and Technology, Shanghai 200237, China Y. Fan Viterbi School of Engineering, University of Southern California, Los Angeles, CA 90089, USA Y. Fan College of International Education, Sichuan International Studies University, Chongqin 400031, China
learn individual task-special transformation for each task. Experimental results on FG-NET, CACD-VS and CALFW datasets show that MDML-BDI achieves satisfactory performance in terms of accuracy and receiver operating characteristic (ROC) curve. Keywords Multi-task . Deep metric learning . Cross-age face verification . Discriminative information
1 Introduction Face recognition has been successfully applied in personal identification, financial activities, security and other fields [1–4]. Face identification, face verification and face authentication are three widely used applications in face recognition. Face recognition searches for a person in a dataset of many face images to solve the problem “Who are you”. Face verification verifies the given two faces are the same person or two different persons, to solve the problem “Is that you”. Face authentication can be viewed as the problem of face verification + “do you have access/permission”, such as the facial authentication payments need no traditional cards, wallets and phones, just the person in the stone. Currently, face verification meets more and more demands of transportation, commerce and public safety. In earlier research, the extracted facial features, posture, illumination and expression are major factors in face verification. Many studies have addressed the face verification with the ex
Data Loading...
