Saccadic trajectory-based identity authentication

The saccadic trajectory is generated by extra-ocular muscles in the eyes, which is a complex mechanism related to brain-driven neural signal. The saccadic trajectory has the characteristics of non-reproducibility and non-contact. In this paper, we propose a saccadic trajectory-based identity authentication method considering that saccadic trajectory can be used as a behavior-based biometric. In this method, we adopt Velocity-Threshold (I-VT) algorithm to extract saccadic trajectories from the whole eye movement data, extract features via wavelet packet transform and authenticate the identity via classifying these features by SVM. In this paper, we verify the proposed method on EMDBv1.0 dataset for horizontal eye movements. We select one subject to be the host and randomly choose another 50 subjects from the remaining 58 subjects as the attackers. We achieve the best performance via optimizing feature selection and the parameter of SVM. The experiment results show that the average accuracy for accepting the host can reach 98.09%, and the average accuracy for rejecting the attackers can reach 99.55%. It demonstrates that the saccadic trajectory-based identity authentication is promising in information security.

[1]  P. Bhattacharya,et al.  Multi-class SVM based iris recognition , 2007, 2007 10th international conference on computer and information technology.

[2]  Daniel L. Silver,et al.  Keystroke and Eye-Tracking Biometrics for User Identification , 2006, IC-AI.

[3]  Weidong Zhou,et al.  Iris recognition based on a novel variation of local binary pattern , 2014, The Visual Computer.

[4]  Tomi Kinnunen,et al.  Towards task-independent person authentication using eye movement signals , 2010, ETRA.

[5]  Wang Yong-qiang.  Feature Extraction and Recognition of Loudspeaker Response Signal Based on Principal Components Analysis , 2009 .

[6]  Yi Yang,et al.  Bi-Level Semantic Representation Analysis for Multimedia Event Detection , 2017, IEEE Transactions on Cybernetics.

[7]  Qinghua Zheng,et al.  Adaptive Unsupervised Feature Selection With Structure Regularization , 2018, IEEE Transactions on Neural Networks and Learning Systems.

[8]  Oleg V. Komogortsev,et al.  Standardization of Automated Analyses of Oculomotor Fixation and Saccadic Behaviors , 2010, IEEE Transactions on Biomedical Engineering.

[9]  R. John Leigh,et al.  A Survey of Eye Movements , 2015 .

[10]  Subhadeep Mukhopadhyay,et al.  LPiTrack: Eye movement pattern recognition algorithm and application to biometric identification , 2018, Machine Learning.

[11]  Jiande Sun,et al.  Eye Recognition With Mixed Convolutional and Residual Network (MiCoRe-Net) , 2018, IEEE Access.

[12]  Cecilia R. Aragon,et al.  Biometric identification via an oculomotor plant mathematical model , 2010, ETRA.

[13]  Nicu Sebe,et al.  The Many Shades of Negativity , 2017, IEEE Transactions on Multimedia.

[14]  Tomi Kinnunen,et al.  Eye-Movements as a Biometric , 2005, SCIA.

[15]  Pawel Kasprowski,et al.  Eye Movements in Biometrics , 2004, ECCV Workshop BioAW.

[16]  Tong Zhang,et al.  Face recognition based on recurrent regression neural network , 2018, Neurocomputing.

[17]  Zheru Chi,et al.  Video-based biometric identification using eye tracking technique , 2012, 2012 IEEE International Conference on Signal Processing, Communication and Computing (ICSPCC 2012).

[18]  Dacheng Tao,et al.  Robust Face Recognition via Multimodal Deep Face Representation , 2015, IEEE Transactions on Multimedia.

[19]  Cecilia R. Aragon,et al.  Biometric authentication via oculomotor plant characteristics , 2012, 2012 5th IAPR International Conference on Biometrics (ICB).

[20]  Ronald R. Coifman,et al.  Entropy-based algorithms for best basis selection , 1992, IEEE Trans. Inf. Theory.