Download presentation
Presentation is loading. Please wait.
1
Face and Gait Biometrics as Surveillance Management System for National Security
ABSTRACT FACE RECOGNITION RESULTS This poster aims to use the biometrics recognition in the surveillance system. Two modalities of biometrics (face and gait) were proposed for the automatic surveillance system. The objectives of this system are identified the outsides of the organization by using the face recognition and to know the terrorist/robber attacks by using the gait recognition. The way of human stands is need to be recognize because the attackers of organization have different way of stand than the normal people. The system cost will be more cheap as compare to the existing surveillance system. This system didn’t need human interaction i.e. self-control and easy to install. While in every big organization, there are already cameras installed to keep the place more security, for that the system will be connected to the same cameras. The novel idea is to recognize the robber/attacker by using the gait biometric because the walk of robber/attacker is different from the normal people. On the basis of face recognition, the proposed system is found to be satisfactory as compare to the existing system. Training Samples Testing Samples Execution time Parallel computing time Recognition Rate 300 features Recognition Rate 600 features 100% 62.25 24.4 97.5% 80% 20% 14.67 6.54 92.5% % 60% 40% 25.18 8.32 % % 36.46 11.74 % % 47.68 14.58 % % Image database Training images Testing images DWT & LBP Feature vectors Feature vector SVM Templet database Matching Result In the pervious table, it shown the recognition rate is 97.5% in 100 % training and 100% testing. Thus, it concludes, even if the same images are used for training and testing, it will still have the possibility to get an error. The average enhancement in the result with 600 features of 1200 samples is 1.8% In the existing system, they have proposed PCA for facial expression database and the result was 89.17% on 300 samples. But, this experiment is achieved % on 300 samples (40% as training and 60% as testing). Feature Extraction Wavelet Transform and Local Binary Pattern Mean and Standard Deviation were computed to obtain the Feature Vector Matching The classification is used to classify the feature vectors. The fit multiclass of Support Vector Machine (SVM) is applied for classifying the feature vectors of the training subjects and it gathers the feature vectors of each subject separately. This classification is available in MATLAB and its function is named as Fitcecoc. Database The Facial Expression Database have been acquired from 160 individual subjects. Where, 15 images have been collected for the face expression. Thus, the total number of facial expressions are 2400 in database. The dimensionality of the Facial Expression Database is 1200×1500 [3]. The AT&T face image database (formerly known as the ORL database) contains a set of face images taken between April 1992 and April 1994 from 40 people [4]. Database Training and Testing Accuracy (%) Expression 40% & 60% 89.44 % ORL 50% & 50% 98.5% INTRODUCTION DWT conclusion The face is the popular way for the humans to recognise each other. The face is the front part of a head from chin to the forehead. Face recognition can be used in surveillance application because the face is one of the few biometric traits which can be recognised by people even at a distance. The Gait is the way of walking. Gait biometric can be used in surveillance application because it can be recognising from distance. The Discrete Wavelet Transform and Local Binary Pattern were used for extracting the features of face. For generating the feature vector, the mean and standard deviation were computed vertically and horizontally. There is a possibility to get an error, even if the same samples used for training are used for testing. The fit multiclass of SVM is used to classify the feature vectors. Since the Facial Expression Database is 2400 images, the normal time of execution is large, so the Parallel Computing Tool is suggested for reducing the time of execution. 223 25 12 120 56 57 20 39 99 Bin2Dec LBP 153 Objectives Identify the outsider of organisation. Recognising the person of interest. Tracing and watching the employee. Recognise the attacker/robber. Bibliography Jaehoon Jung, Inhye Yoon and Joonki Paik “Object Occlusion Detection Using Automatic Camera Calibration for a Wide-Area Video Surveillance System”, Sensors MDPI , June 2016 My Self and my Guide “Extraction of Palmprint Texture Features using Combined DWT-DCT and Local Binary Pattern”, 2nd IEEE International Conference on Next Generation Computing Technologies (NGCT-2016), Dehradun, Indai, October 2016. Dinesh N. Satange “NC-FACE Database for Face and Facial Expression Recognition” International Journal of Pure and Applies in Research in Engineering and Technology (IJPRET) Volume 3 (9) , 2015 Olivetti Research Laboratory (ORL) face database. Is available in the following link
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.