Thuc Dinh Nguyen

Adaptive Cross-Device Gait Recognition Using a Mobile Accelerometer

Mobile authentication/identification has grown into a priority issue nowadays because of its existing outdated mechanisms, such as PINs or passwords. In this paper, we introduce gait recognition by using a mobile accelerometer as not only effective but also as an implicit identification model. Unlike previous works, the gait recognition only performs well with a particular mobile specification (e.g., a fixed sampling rate). Our work focuses on constructing a unique adaptive mechanism that could be independently deployed with the specification of mobile devices. To do this, the impact of the sampling rate on the preprocessing steps, such as noise elimination, data segmentation, and feature extraction, is examined in depth. Moreover, the degrees of agreement between the gait features that were extracted from two different mobiles, including both the Average Error Rate (AER) and Intra-class Correlation Coefficients (ICC), are assessed to evaluate the possibility of constructing a device-independent mechanism. We achieved the classification accuracy approximately for both devices, which showed that it is feasible and reliable to construct adaptive cross-device gait recognition on a mobile phone.

Real time traffic sign detection using color and shape-based features

This paper presents a new approach for color detection and segmentation based on Support Vector Machine (SVM) to retrieve candidate regions of traffic signs in real-time video processing. Instead of processing on each pixel, this approach utilizes a block of pixels as an input vector of SVM for color classification, where the dimension of each vector can be extended by a group of neighboring pixels. This helps to handle the diversification of data on both training and testing samples. After that, Hough transform and contour detection are applied to verify the candidate regions by detecting shapes of circle and triangle. The experimental results are highly accurate and robust for our testing database, where samples are recorded on various states of environment.

A Lightweight Gait Authentication on Mobile Phone Regardless of Installation Error

In this paper, we propose a novel gait authentication mechanism by mining sensor resources on mobile phone. Unlike previous works, both built-in accelerometer and magnetometer are used to handle mobile installation issues, including but not limited to disorientation, and misplacement errors. The authentication performance is improved by executing deep examination at pre-processing steps. A novel and effective segmentation algorithm is also provided to segment signal into separate gait cycles with perfect accuracy. Subsequently, features are then extracted on both time and frequency domains. We aim to construct a lightweight but high reliable model; hence feature subsets selection algorithms are applied to optimize the dimension of the feature vectors as well as the processing time of classification tasks. Afterward, the optimal feature vector is classified using SVM with RBF kernel. Since there is no public dataset in this field to evaluate fairly the effectiveness of our mechanism, a realistic dataset containing the influence of mobile installation errors and footgear is also constructed with the participation of 38 volunteers (28 males, 10 females). We achieved the accuracy approximately 94.93% under identification mode, the FMR, FNMR of 0%, 3.89% and processing time of less than 4 seconds under authentication mode.

Efficiently Preserving Data Privacy Range Queries in Two-Tiered Wireless Sensor Networks

Wireless Sensor Networks (WSNs) gain popularity these days. WSNs have been one of key technologies for the future with broad applications from the military to everyday life. There are two kinds of WSNs model models with sensors for sensing data and a server for receiving and processing queries from users, and models with special additional nodes, their common names are storage nodes, between the sensors and the sink. Among the latter type, a two-tiered model has been widely adopted because of its storage and energy saving benefits for weak sensors, as proved by the advent of commercial storage node products such as Stargate and RISE. In this paper, we consider two-tiered WSNs model that consists of sensor nodes receiving data from the environment and transferred to storage nodes, the queries from the server will be done through these storage nodes. In this model, the problem is how to ensure privacy of data at the storage nodes, which is mostly located in the hostile area. Data encryption is the popular solution to make the unreadable for storage node. However, if the server needs data in the range, then storage node must read the data to reply servers query. In order to solve this problem, we proposed to use a d-disjunct matrix, an order-preserving function and a permutation function that encode data of each sensor. An order-preserving function is used to query. Simultaneously, the permutation function and d-disjunct matrix will be used to encrypt data and determine the exact frequency of any value. With this approach, we have proved that our solution guarantees to preserve data-privacy while allowing the range query efficiently. In particular, it just takes less than 60 seconds when we encrypt or query less than 200000 values.

An Efficient Pseudo Inverse Matrix - Based Solution for Secure Auditing

Database outsourcing is emerging as an important trend in which the data owners ship their data to external service provider. Since a service provider is not fully trusted, security and privacy are significant issues. They can be referred to as data confidentiality, user and data privacy, query assurance and secure auditing. Among them, secure auditing takes an important role to have a secure database outsourcing model. While audit log provides investigator information to assess past system activity, it often contains sensitive information. We need to protect it from unauthorized parties while making it still efficiently searchable by authorized ones. In this paper, we propose a novel searchable encryption scheme, named PEKS-PM, based on pseudo inverse matrix, which enables a trusted party to give the service providers server the ability to test whether a given keyword appears in log records but the server learns nothing about the keyword and the log content. Our encryption scheme is based mainly on linear matrix operations therefore the complexity is low. We prove that our proposed encryption scheme is secure and efficient, comparing with an encryption scheme based on the Decisional Diffie-Hellman (DDH) assumption.

Robust Message Hiding for QR Code

Response Code (QR code) is widely used in daily life in recent years because it has high capacity encoding of data, damage resistance, fast decoding and other good characteristics. Since it is popular, people can use it to transmit secret information without inspection. The development of steganography in QR code lead to many problems arising. How to keep the original content of QR code and embed secret information into it are the two main challenges. Hiding secret information based on bit technique is so fragile to modification attack. If an attacker change any bit of hidden bits, it is impossible to recover the secret information. In this paper, we proposed a scheme based on Reed-Solomon codes and List Decoding to overcome this problem. We also conduct our solution by analyzing the complexity, security, and experiment.

Quasi-inverse Based Cryptography

We are interested in monoids and its applications. If every element x in a monoid has a quasi-inverse y in the sense of von Neumann, that satisfies x ·y ·x = x and y ·x ·y = y, the monoid is regular. Our purpose is to use regular monoids to build two abstract algebraic public key cryptosystems: key exchange protocol and public key encryption with keyword search scheme. In addition to illustrating how the two cryptosystems work, we provide instances of these abstract algebraic models and analyse them in terms of cryptanalysis security.

On the instability of sensor orientation in gait verification on mobile phone

Authentication schemes using tokens or biometric modalities have been proposed to ameliorate the security strength on mobile devices. However, the existing approaches are obtrusive since the user is required to perform explicit gestures in order to be authenticated. While the gait signal captured by inertial sensors is understood to be a reliable profile for effective implicit authentication, recent studies have been conducted in ideal conditions and might therefore be inapplicable in the real mobile context. Particularly, the acquiring sensor is always fixed to a specific position and orientation. This paper mainly focuses on addressing the instability of sensors orientation which mostly happens in the reality. A flexible solution taking advantages of available sensors on mobile devices which can help to handle this problem is presented. Moreover, a novel gait recognition method utilizes statistical analysis and supervised learning to adapt itself to the instability of the biometric gait under various circumstances is also proposed. By adopting PCA+SVM to construct the gait model, the proposed method outperformed other state-of-the-art studies, with an equal error rate of 2.45% and accuracy rate of 99.14% in terms of the verification and identification aspects being achieved, respectively.

A variant of non-adaptive group testing and its application in pay-television via internet

In non-adaptive group testing (NAGT), the time for decoding is a crucial problem. Given an unknown string x∈{0, 1}N with at most d ones, the problem is how to determine xi=1 using as few tests as possible so that x can be decoded as fast as possible. A NAGT can be represented by a t ×N matrix. Although we do not know x, this matrix, which is called d-disjunct matrix, can reconstruct it exactly. In this paper, we consider a general problem, in which x is an array of N non-negative integer elements and has up to d positive integers. From nonrandom construction, we prove that we can decode a d-disjunct matrix, which is built from [n, k]q-Reed-Solomon codes and identity matrix Iq, and recover x defined above in poly(d) ·t log2t+O(d3n log(d logN)) with t=O(d2log2N). We also discuss this problem when x contains negative integer elements. Pay-Television internet-based can be applied these results directly. Since the number of customers is very large, our system must be prevented from illegal buyers. This problem is called traitor tracing. To the best of our knowledge, this is the first result that raises a variant of NAGT and gets how to trace traitors without using probability.

Early detection for networking anomalies using non-adaptive group testing

Monitoring operation of thousands of servers and network devices on big data centers or server farms are very important roles for administrators to ensure that it well-operates, early detection of anomalies, fast errors correcting and decreasing discontinuous network. Network monitoring system detects anomalies such as attacks, states of hosts or services, resources. The aim of this is recognizes network faults and attacks quickly. A method widely used for almost network monitoring systems is setup agents on servers, network devices and then establishing connections between them and monitoring servers using some protocols such as ICMP and SNMP. These servers send periodic requests to agents to get reports or agents send traps to network monitoring servers. With this method, it must long time to alert. We propose a solution to fast detecting some anomalies such as servers or devices operate with high frequency (called “hot IPs”) and low frequency (called “low IPs”) and it works independently to early warning for these anomalies using non-adaptive group testing method. In particular, if dealing with up to 260,000 IPs, we can detect up to 31 hot and low IPs within 2.5 minutes.