Biblio

Compressive sensing (CS) is beneficial for unmanned reconnaissance systems to obtain high-quality images with limited resources. The existing prediction methods for block-based compressive sensing of images can be regarded as the particular coefficients of weighted predictive coding. To find better prediction coefficients for BCS, this paper proposes two weighted prediction methods. The first method converts the prediction model of measurements into a prediction model of image blocks. The prediction weights are obtained by training the prediction model of image blocks offline, which avoiding the influence of the sampling rates on the prediction model of measurements. Another method is to calculate the prediction coefficients adaptively based on the average energy of measurements, which can adjust the weights based on the measurements. Compared with existing methods, the proposed prediction methods for BCS of images can further improve the reconstruction image quality.

The rapid proliferation of biometrics has led to growing concerns about the security and privacy of the biometric data (template). A biometric uniquely identifies an individual and unlike passwords, it cannot be revoked or replaced since it is unique and fixed for every individual. To address this problem, many biometric template protection methods using fully homomorphic encryption have been proposed. But, most of them (i) are computationally expensive and practically infeasible (ii) do not support operations over real valued biometric feature vectors without quantization (iii) do not support packing of real valued feature vectors into a ciphertext (iv) require multi-shot enrollment of users for improved matching performance. To address these limitations, we propose a secure and privacy preserving method for biometric template protection using fully homomorphic encryption. The proposed method is computationally efficient and practically feasible, supports operations over real valued feature vectors without quantization and supports packing of real valued feature vectors into a single ciphertext. In addition, the proposed method enrolls the users using one-shot enrollment. To evaluate the proposed method, we use three face datasets namely LFW, FEI and Georgia tech face dataset. The encrypted face template (for 128 dimensional feature vector) requires 32.8 KB of memory space and it takes 2.83 milliseconds to match a pair of encrypted templates. The proposed method improves the matching performance by 3 % when compared to state-of-the-art, while providing high template security.

Securing wireless communication is significant for privacy and confidentiality of sensing data in Cyber-Physical Systems (CPS). However, due to broadcast nature of radio channels, disseminating sensory data is vulnerable to eavesdropping and message modification. Generating secret keys by extracting the shared randomness in a wireless fading channel is a promising way to improve the communication security. In this poster, we present a novel secret key generation protocol for securing real-time data dissemination in CPS, where the sensor nodes cooperatively generate a shared key by estimating the quantized fading channel randomness. A 2-hop wireless sensor network testbed is built and preliminary experimental results show that the quantization intervals and distance between the nodes lead to a secret bit mismatch.

This work studies the optimization of spectrum pooling for the downlink of a multi-tenant Cloud Radio Access Network (C-RAN) system in the presence of inter-tenant privacy constraints. The spectrum available for downlink transmission is partitioned into private and shared subbands, and the participating operators cooperate to serve the user equipments (UEs) on the shared subband. The network of each operator consists of a cloud processor (CP) that is connected to proprietary radio units (RUs) by means of finite-capacity fronthaul links. In order to enable inter-operator cooperation, the CPs of the participating operators are also connected by finite-capacity backhaul links. Inter-operator cooperation may hence result in loss of privacy. The problem of optimizing the bandwidth allocation, precoding, and fronthaul/backhaul compression strategies is tackled under constraints on backhaul and fronthaul capacity, as well as on per-RU transmit power and inter-onerator privacy.

Chaotic dynamics is widely used to design pseudo-random number generators and for other applications, such as secure communications and encryption. This paper aims to study the dynamics of the discrete-time chaotic maps in the digital (i.e., finite-precision) domain. Differing from the traditional approaches treating a digital chaotic map as a black box with different explanations according to the test results of the output, the dynamical properties of such chaotic maps are first explored with a fixed-point arithmetic, using the Logistic map and the Tent map as two representative examples, from a new perspective with the corresponding state-mapping networks (SMNs). In an SMN, every possible value in the digital domain is considered as a node and the mapping relationship between any pair of nodes is a directed edge. The scale-free properties of the Logistic map's SMN are proved. The analytic results are further extended to the scenario of floating-point arithmetic and for other chaotic maps. Understanding the network structure of a chaotic map's SMN in digital computers can facilitate counteracting the undesirable degeneration of chaotic dynamics in finite-precision domains, also helping to classify and improve the randomness of pseudo-random number sequences generated by iterating the chaotic maps.

The storage efficiency of hash codes and their application in the fast approximate nearest neighbor search, along with the explosion in the size of available labeled image datasets caused an intensive interest in developing learning based hash algorithms recently. In this paper, we present a learning based hash algorithm that utilize ordinal information of feature vectors. We have proposed a novel mathematically differentiable approximation of argmax function for this hash algorithm. It has enabled seamless integration of hash function with deep neural network architecture which can exploit the rich feature vectors generated by convolutional neural networks. We have also proposed a loss function for the case that the hash code is not binary and its entries are digits of arbitrary k-ary base. The resultant model comprised of feature vector generation and hashing layer is amenable to end-to-end training using gradient descent methods. In contrast to the majority of current hashing algorithms that are either not learning based or use hand-crafted feature vectors as input, simultaneous training of the components of our system results in better optimization. Extensive evaluations on NUS-WIDE, CIFAR-10 and MIRFlickr benchmarks show that the proposed algorithm outperforms state-of-art and classical data agnostic, unsupervised and supervised hashing methods by 2.6% to 19.8% mean average precision under various settings.

Digital video has various types of entities, which are utilized as embedding domains to hide messages in steganography. However, nearly all video steganography uses only one type of embedding domain, resulting in limited embedding capacity and potential security risks. In this paper, we firstly propose to embed in multi-domains for video steganography by combining partition modes (PMs) and motion vectors (MVs). The multi-domain embedding (MDE) aims to spread the modifications to different embedding domains for achieving higher undetectability. The key issue of MDE is the interactions of entities across domains. To this end, we design two MDE strategies, which hide data in PM domain and MV domain by sequential embedding and simultaneous embedding respectively. These two strategies can be applied to existing steganography within a distortion-minimization framework. Experiments show that the MDE strategies achieve a significant improvement in security performance against targeted steganalysis and fusion based steganalysis.

We propose a new key sharing protocol executed through any constant parameter noiseless public channel (as Internet itself) without any cryptographic assumptions and protocol restrictions on SNR in the eavesdropper channels. This protocol is based on extraction by legitimate users of eigenvalues from randomly generated matrices. A similar protocol was proposed recently by G. Qin and Z. Ding. But we prove that, in fact, this protocol is insecure and we modify it to be both reliable and secure using artificial noise and privacy amplification procedure. Results of simulation prove these statements.

Model compression is considered to be an effective way to reduce the implementation cost of deep neural networks (DNNs) while maintaining the inference accuracy. Many recent studies have developed efficient model compression algorithms and implementations in accelerators on various devices. Protecting integrity of DNN inference against fault attacks is important for diverse deep learning enabled applications. However, there has been little research investigating the fault resilience of DNNs and the impact of model compression on fault tolerance. In this work, we consider faults on different data types and develop a simulation framework for understanding the fault resiliency of compressed DNN models as compared to uncompressed models. We perform our experiments on two common DNNs, LeNet-5 and VGG16, and evaluate their fault resiliency with different types of compression. The results show that binary quantization can effectively increase the fault resilience of DNN models by 10000x for both LeNet5 and VGG16. Finally, we propose software and hardware mitigation techniques to increase the fault resiliency of DNN models.

We investigate how to protect features corresponding to personal information using homomorphic encryption when matching people in several camera views. Homomorphic encryption can compute a distance between features without decryption. Thus, our method is able to use a computing server on a public network while protecting personal information. To apply homomorphic encryption, our method uses linear quantization to represent each element of the feature as integers. Experimental results show that there is no significant difference in the accuracy of person re-identification with or without homomorphic encryption and linear quantization.

Key derivation from the physical layer features of the communication channels is a promising approach which can help the key management and security enhancement in communication networks. In this paper, we consider a key generation technique that quantizes the received signal phase to obtain the secret keys. We then study the effect of a jamming attack on this system. The jammer is an active attacker that tries to make a disturbance in the key derivation procedure and changes the phase of the received signal by transmitting an adversary signal. We evaluate the effect of jamming on the security performance of the system and show the ways to improve this performance. Our numerical results show that more phase quantization regions limit the probability of successful attacks.

In this paper, we have considered the broadcasting of a memoryless bivariate Gaussian source over a Gaussian broadcast channel with respect to bandwidth compression. We have analysed the performance of a hybrid digital-analog (HDA) coding system in combination with joint source channel coding (JSCC) to measure the distortion regions. The transmission advantages due to the combination of both the analog and digital techniques, a class of HDA schemes that yields better performance in distortion is discussed. The performance of source and channel coding for the possible better outcome of the system is measured by employing Wyner-Ziv and Costa coding. In our model, we have considered the upper layer to be a combination of a hybrid layer in the sense of both the analog and digital processing is done. This is executed in presence of quantization error and performance of the system is measured with two conditions: 1) HDA scheme with quantization scaling factor α = 0, i.e. the input of the channel have only the analog information which is considered as the scaled quantization error βS 2) The analog information from the first layer S is suppressed by setting error scaling factor β = 0 and 3) Inclusion of recursive mode with JSCC in each of the three layers for the possible better outcome is considered here.

The rapid development of Internet has resulted in massive information overloading recently. These information is usually represented by high-dimensional feature vectors in many related applications such as recognition, classification and retrieval. These applications usually need efficient indexing and search methods for such large-scale and high-dimensional database, which typically is a challenging task. Some efforts have been made and solved this problem to some extent. However, most of them are implemented in a single machine, which is not suitable to handle large-scale database.In this paper, we present a novel data index structure and nearest neighbor search algorithm implemented on Apache Spark. We impose a grid on the database and index data by non-empty grid cells. This grid-based index structure is simple and easy to be implemented in parallel. Moreover, we propose to build a scalable KNN graph on the grids, which increase the efficiency of this index structure by a low cost in parallel implementation. Finally, experiments are conducted in both public databases and synthetic databases, showing that the proposed methods achieve overall high performance in both efficiency and accuracy.

Binary embedding is an effective way for nearest neighbor (NN) search as binary code is storage efficient and fast to compute. It tries to convert real-value signatures into binary codes while preserving similarity of the original data. However, it greatly decreases the discriminability of original signatures due to the huge loss of information. In this paper, we propose a novel method double-bit quantization and weighting (DBQW) to solve the problem by mapping each dimension to double-bit binary code and assigning different weights according to their spatial relationship. The proposed method is applicable to a wide variety of embedding techniques, such as SH, PCA-ITQ and PCA-RR. Experimental comparisons on two datasets show that DBQW for NN search can achieve remarkable improvements in query accuracy compared to original binary embedding methods.

We propose a multi-level CSI quantization and key reconciliation scheme for physical layer security. The noisy wireless channel estimates obtained by the users first run through a transformation, prior to the quantization step. This enables the definition of guard bands around the quantization boundaries, tailored for a specific efficiency and not compromising the uniformity required at the output of the quantizer. Our construction results in an better key disagreement and initial key generation rate trade-off when compared to other level-crossing quantization methods.

Advances in virtual reality have generated substantial interest in accurately reproducing and storing spatial audio in the higher order ambisonics (HOA) representation, given its rendering flexibility. Recent standardization for HOA compression adopted a framework wherein HOA data are decomposed into principal components that are then encoded by standard audio coding, i.e., frequency domain quantization and entropy coding to exploit psychoacoustic redundancy. A noted shortcoming of this approach is the occasional mismatch in principal components across blocks, and the resulting suboptimal transitions in the data fed to the audio coder. Instead, we propose a framework where singular value decomposition (SVD) is performed after transformation to the frequency domain via the modified discrete cosine transform (MDCT). This framework not only ensures smooth transition across blocks, but also enables frequency dependent SVD for better energy compaction. Moreover, we introduce a novel noise substitution technique to compensate for suppressed ambient energy in discarded higher order ambisonics channels, which significantly enhances the perceptual quality of the reconstructed HOA signal. Objective and subjective evaluation results provide evidence for the effectiveness of the proposed framework in terms of both higher compression gains and better perceptual quality, compared to existing methods.

This paper reports a research work on how to transmit a secured image data using Discrete Wavelet Transform (DWT) in combination of Advanced Encryption Standard (AES) with low power and high speed. This can have better quality secured image with reduced latency and improved throughput. A combined model of DWT and AES technique help in achieving higher compression ratio and simultaneously it provides high security while transmitting an image over the channels. The lifting scheme algorithm is realized using a single and serialized DT processor to compute up to 3-levels of decomposition for improving speed and security. An ASIC circuit is designed using RTL-GDSII to simulate proposed technique using 65 nm CMOS Technology. The ASIC circuit is implemented on an average area of about 0.76 mm2 and the power consumption is estimated in the range of 10.7-19.7 mW at a frequency of 333 MHz which is faster compared to other similar research work reported.

Nearest neighbor search is a fundamental problem in various research fields like machine learning, data mining and pattern recognition. Recently, hashing-based approaches, for example, locality sensitive hashing (LSH), are proved to be effective for scalable high dimensional nearest neighbor search. Many hashing algorithms found their theoretic root in random projection. Since these algorithms generate the hash tables (projections) randomly, a large number of hash tables (i.e., long codewords) are required in order to achieve both high precision and recall. To address this limitation, we propose a novel hashing algorithm called density sensitive hashing (DSH) in this paper. DSH can be regarded as an extension of LSH. By exploring the geometric structure of the data, DSH avoids the purely random projections selection and uses those projective functions which best agree with the distribution of the data. Extensive experimental results on real-world data sets have shown that the proposed method achieves better performance compared to the state-of-the-art hashing approaches.

This paper presents a novel design of content fingerprints based on maximization of the mutual information across the distortion channel. We use the information bottleneck method to optimize the filters and quantizers that generate these fingerprints. A greedy optimization scheme is used to select filters from a dictionary and allocate fingerprint bits. We test the performance of this method for audio fingerprinting and show substantial improvements over existing learning based fingerprints.