Transactions on Computational Science XXX

Recent research in the areas of computer graphics, virtual reality and cyberworlds is increasingly concerned with the security applications. In the area of cyberworlds, human features such as faces, hair, walking patterns, voice, behaviour and a manner of communications are being simulated and studied. Moreover, research into social online interactions, and an effort to mimic those interactions and appearances through virtual humans and robots has become abundant. This position paper discusses the state-of-the-art research in the fields of biometric recognition, multi-modal and cancelable biometrics, artificial biometrics and social behavioral studies conducted in the Biometric Technologies laboratory at the University of Calgary, Canada. Social behavioral pattern analysis is the emerging domain in the biometric recognition; the idea behind it is to extract behavioral clues from the everyday human interactions. Activity-related biometric authentication provides an unobtrusive and natural alternative for physiological biometric that can exploit everyday life activities involving interaction with objects or people for extracting biometric signature. Recent research demonstrated that it is possible to extract the behavioral traits not only from traditional behavioral identifiers, such as voice, gait or signature, but also from an online interactions of users or an editing behavior of Wikipedia article authors. An overview of this emerging areas of research that brings virtual worlds modeling and biometric security recognition fields together has been presented at DRDC 2017 National Defense and Security Workshop, Canada. This Editor-in-Chief Position paper is based on the results discussed there as well as on the report submitted to DRDC following the workshop (Gavrilova 2017).

Automated face recognition is commonly used for security reinforcement and identity verification purposes. While significant advancement has been made in this domain, modern surveillance techniques are still dependent on variations in pose, orientation of the facial images, difference in the illumination, occlusion, etc. Therefore, face recognition or identification in uncontrolled situations has become an important research topic. In this paper, we propose a new face recognition technique that takes into account partial occlusion, while still accurately identifying the user. The occluded facial areas are detected from the Kinect depth images by extracting features using Uniform Local Binary Pattern (LBP). For localizing occluded regions from the Kinect depth images, a threshold based approach is used to identify the areas close to the camera. The recognition system will discard the occluded regions of the facial images and match only the non-occluded facial part with the gallery of images to find the best possible match. The performance of the recognition system has been evaluated on EUROKOM Kinect face database containing different types of occluded and non-occluded faces with neutral expressions. Experimental results show that the proposed method improves the recognition rate by 4.8% and 5.7% for occlusion by hand and occlusion by paper, respectively.

This paper proposes a new method of transferring style between images by considering scene matching between the source image and the target image. Artists often employ different colors and brushwork for individual subjects. Likewise, the connections between various subjects in a work also affect the colors and brushwork used. Our method begins with input images, searches an example database for paintings with scenes similar to that in the input image, and transfers the color and brushwork of the paintings to the corresponding target images to generate painterly images that reflect specific styles. Our method applies a GIST approach to the process of searching for paintings with similar scenes before performing style transfers. The spatial correspondence between the source image and the target image is also used to ensure close correlation between various elements in order to reproduce styles faithfully.

Skyline query has been investigated extensively in many fields recently. One of the interesting ones is the spatial skyline query that retrieves those points of P not dominated by any other point in P with respect to their derived spatial attributes. However, being point-based, this kind of query method is not suitable for privacy protection. In this paper, we introduce the privacy-preserved spatial skyline query where the distances calculated between the query points and the objects change from ‘point to point’ to ‘region to point’. It is the first effort to process relative skyline queries based on a ‘region to point’ way. Accordingly, we proposed three approaches: BC, VRS² and NVRS². While BC is a straightforward method, VRS² and NVRS² manipulate the properties of Voronoi diagram and Network Voronoi diagram for the Euclidean space and road networks situations respectively. Furthermore, with respect to the changes of query conditions, another two algorithms DPJA and DPDA to dynamically update the results are proposed so that the heavy re-calculation could be avoided. Our empirical experiments show that our approaches have good performance in retrieving the skyline points of privacy-preserved spatial skyline query.

Cybersecurity is an important and challenging issue faced by the governments, financial institutions and ordinary citizens alike. Secure identification is needed for accessing confidential personal information, online bank transactions, people’s social networks etc. Brain signal electroencephalogram (EEG) can play a vital role in ensuring security as it is non-vulnerable and very difficult to forge. In this article, we develop an EEG based biometric security system. The purpose of this research is to find the relationship between thinking capability and person identification accuracy by comparison analyzing of overt and covert mental stimuli of brain signal. The Discrete Wavelet Transform (DWT) is used to extract different significant features which separate Alpha, Beta and Theta band of frequencies of the EEG signal. Extracted EEG features of different bands and their combinations such as alpha-beta, alpha-theta, theta-beta, alpha-beta-theta are classified using an artificial neural network (ANN) trained with the back propagation (BP) algorithm. Another classifier K-nearest neighbors (KNN) is used to verify the results of this experiment. Both classification results show that alpha band has a higher convergence rate than other bands, beta and theta, for the overt EEG signal. From overt mental stimuli, we also discover that individual band provides better performance than band combination. So, we have applied Back Propagation (BP) algorithm at individual band of various features of covert mental stimuli and obtained the accuracy 73.1%, 78.1% and 74.4% for alpha, beta and theta band respectively. By comparing the analysis of overt and covert mental stimuli, the overt brain signal shows better performance. Finally, we conclude that the relationship between thinking capability and person identification accuracy is inversely proportional. The results of this study are expected to be helpful for future research by using various thinking capability brain signals based biometric approaches.

In this study, a Man-machine Finger-guessing game is designed based on the IntelliSense and Man-machine coordination mechanism of hand gesture. The image sequence is obtained by the Kinect and the human hand is extracted using segmentation and skin color modeling. The proposed SCDDF (Shape Context Density Distribution Feature), which combined DDF (Density Distribution Feature) algorithm and shape context recognition algorithm, is used to extract gesture identity. Gestures are finally identified by registering with templates in the pre-established gesture library. Furthermore, we proposed a new human-computer cooperative mechanism, including two points: (1) The virtual interface is used to control the ‘Midas Touch problem’. (2) The whole game is more natural and smooth. In the aspect of gesture recognition, we combined DDF algorithm and shape context recognition algorithm, and proposed the SCDDF algorithm. The new algorithm improved recognition rate by 14.3% compared with DDF algorithm.