3D Face Detection and Recognition under Occlusion Dr.V.Ramaswamy 1 , Parashuram Baraki 2 1 Research Guide, Jain University, Bangalore, 2 Doctoral Student, Jain University, Bangalore & Asst.Professor , CS&E, Dept, GM Institute of Technology, Davanagere Abstract: Human face is undoubtedly one of the most deserving biometric an aspect for which automatic authentication is very vital. Three-dimensional face recognition exploits facial surface information. Compared to 2D face recognition, 3D exhibits good robustness and high fake resistance. This makes it a good candidate to be used in high security areas. Face recognition is sensitive to aspects such as variation, illumination and occlusion. However, dealing with occlusions covering the facial surface is a great challenge which should be handled in such a way as to enable applicability to fully automatic security systems. In this paper, we propose an automatic 3D face detection and recognition system which is robust to occlusions. We basically consider two problems namely occlusion handling for surface registration and recognition with missing data. Face can be detected by using Viola Jones technique. Registration scheme makes use of Iterative Closest Point technique. Principal Component Analysis is being used to deal with recognition of face. Experimental results confirm that proposed techniques offer an occlusion robust face detection and recognition system. Index Terms- 3D face, Iterative closest point, viola Jones, Principal Component Analysis I. INTRODUCTION Biometric is an automated method of recognizing a person based on a physiological or behavioural characteristic. Biometric technologies are becoming the foundation of an extensive array of highly secure identification and personal verification solutions. As the level of security breaches and transaction fraud increases, the need for highly secure identification and personal verification technologies is becoming very relevant. Biometric based solutions provide confidential financial transactions and personal data privacy. From a biometric point of view, human face is widely preferred because of several advantages: Due to its contactless acquisition, it is well accepted among users. Furthermore, its applicability to non-cooperative scenarios makes it suitable for a range of applications such as surveillance systems. The factors that degrade the performance of a face recognizer include presence of illumination differences, in-depth pose variations, facial expression variations besides the presence of occlusions. Two kinds of errors are being considered by a biometric system. They are false rejection – a legitimate user is rejected and false acceptance – an impostor is accepted as a legitimate user. Face detection and recognition requires a digital camera to develop a facial image of the user for authentication. It is evident that identification is technically more challenging and costly. Identification accuracy generally decreases as the size of the database grows. Before the user can be successfully recognized by the system, he/she must be registered with the biometric system. User’s biometric data is captured, processed and stored. As the quality of this stored biometric data is very crucial for further authentications, often several (usually 3 or 5) biometric samples are used to create user’s master template. The process of the user’s registration with the biometric system is called enrolment. II. RELATED WORK Developments in 3-D sensor technologies have increased interest in 3-D face recognition. In [1], it is shown that by using 3-D face, it is possible to obtain competitive results when compared to other areas such as iris and high-resolution 2-D facial images. A thorough survey of previously proposed 3-D face recognition systems can be found in [1]. In this section, we focus on the recent face recognition approaches which deal with realistic occlusion variations in 3-D. In the studies using 3-D facial data, only a few consider facial occlusion detection, removal, restoration and missing data handling. In [11], author has proposed a method to detect occlusions by analyzing the difference between an original face and its approximation using Eigen face approach. The regions detected as occlusions are removed and the locations of the missing parts are employed in the restoration process which is handled by Gappy PCA [7]. In [3], a part-based method is proposed in which facial regions are aligned independently to average regional models. Regional division scheme is also employed in the classification stage where various regional classification results are fused with different fusion techniques. Experimental results indicate performance improvement by part-based system both for expression and occlusion variations. In [6], authors have proposed a nose- based registration scheme for better handling of occluded faces. Curvature information is utilized for automatic detection of nose area and an average nose model through Iterative Closest Point (ICP) algorithm is used for fine alignment. On the registered surfaces, occlusions are detected by analyzing the difference from the average face model. Occlusion removal takes place using a modified version of PCA method. Restored faces are classified using different local masks and multiple classifiers are fused for final identity estimation. V.Ramaswamy et al, / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 5 (6) , 2014, 7972-7976 www.ijcsit.com 7972