This paper proposes a novel method for object tracking by combining local feature and global template-based methods. The proposed algorithm consists of two stages from coarse to fine. The first stage applies on-line classifiers to match the corresponding keypoints between the input frame and the reference frame. Thus a rough motion parameter can be estimated using RANSAC. The second stage employs kernel-based global representation in successive frames to refine the motion parameter. In addition, we use the kernel weight obtained during the second stage to guide the on-line learning process of the keypoints' description. Experimental results demonstrate the effectiveness of the proposed technique.

Blind source separation (BSS) techniques are required for various signal decomposing issues. Independent component analysis (ICA), assuming only a statistical independence among stochastic source signals, is one of the most useful BSS tools because it does not need a priori information on each source. However, there are many requirements for decomposing multiple deterministic signals such as complex sinusoidal signals with different frequencies. These requirements may include pulse compression or clutter rejection. It has been theoretically shown that an ICA algorithm based on maximizing non-Gaussianity successfully decomposes such deterministic signals. However, this ICA algorithm does not maintain a sufficient separation performance when the frequency difference of the sinusoidal waves becomes less than a nominal frequency resolution. To solve this problem, this paper proposes a super-resolution algorithm for complex sinusoidal signals by extending the maximum likelihood ICA, where the probability density function (PDF) of a complex sinusoidal signal is exploited as a priori knowledge, in which the PDF of the signal amplitude is approximated as a Gaussian distribution with an extremely small standard deviation. Furthermore, we introduce an optimization process for this standard deviation to avoid divergence in updating the reconstruction matrix. Numerical simulations verify that our proposed algorithm remarkably enhances the separation performance compared to the conventional one, and accomplishes a super-resolution separation even in noisy situations.

The discharge properties and chemical surface stability of CeO2 containing Sr (CeSrO) as the candidate for high-γ protective layer of noble plasma display panels (PDPs) are characterized. CeSrO films have superior chemical stability, because of the decrease in reactiveness on surface due to their fluorite structure. The discharge voltage is 50 V lower than that of MgO films for a pure discharge gas of Ne/Xe = 85/15 at 60 kPa. However the topmost surface, monolayer, of the CeSrO film relevant to the discharge property is hardly recovered from the damage by CO2 impurity in discharge gas. We can expect that by pumping down to a sufficiently low CO2 partial pressure (lower than 1 10-3 Pa), PDP panels with very high efficiency are realized with CeSrO protective layer.

High-speed IP address lookup with fast prefix update is essential for designing wire-speed packet forwarding routers. The developments of optical fiber and 100 Gbps interface technologies have placed IP address lookup as the major bottleneck of high performance networks. In this paper, we propose a novel structure named Compressed Multi-way Prefix Tree (CMPT) based on B+ tree to perform dynamic and scalable high-speed IP address lookup. Our contributions are to design a practical structure for high-speed IP address lookup suitable for both IPv4 and IPv6 addresses, and to develop efficient algorithms for dynamic prefix insertion and deletion. By investigating the relationships among routing prefixes, we arrange independent prefixes as the search indexes on internal nodes of CMPT, and by leveraging a nested prefix compression technique, we encode all the routing prefixes on the leaf nodes. For any IP address, the longest prefix matching can be made at leaf nodes without backtracking. For a forwarding table with u independent prefixes, CMPT requires O(logmu) search time and O(mlogmu) dynamic insert and delete time. Performance evaluations using real life IPv4 forwarding tables show promising gains in lookup and dynamic update speeds compared with the existing B-tree structures.

The authors have been investigating degradation process of Au plated slip ring and Ag-Pd-Cu brush system. In almost all cases the lifetime of the sliding system ends, when Au plating layer is worn out, the ring surface is oxidized to be black in color and contact resistance becomes very high. Further, the lifetime is very short without lubricant. So, the lubricant is very effective to make the lifetime longer. However, even with lubricant the lifetime is varied from about 1000 hours to almost 7000 hours in the past experiments. It is an important issue how the lubricant works on the lifetime of the system. In this paper the effect of lubricant on the degradation process of contact resistance is focused on. In the past tests the lubricant is supplied only once before the test. In this test the lubricant is regularly supplied almost every 900 operation hours. Consequently, the operation more than 8000 hours is realized, which is the longest among tests so far. In addition the contact voltage drop increase gradually until 2600 hours and after that it stays almost constant around 70 mV. According to the Element Analysis after the test the Ni base plating layer is totally exposed in many tracks. It means that the Au plating layer is gradually worn out probably at the stage of increasing voltage drop. In the previous tests the lifetime ended even when the Ni plating layer remained. So, the reason of long operation in this test is guessed to be that the lubricant not only decreases wear of ring and brush, but also suppresses oxidation of the Ni layer.

Recently, dynamic power supply voltage techniques, such as an Envelope Elimination and Restoration power amplifier (EER-PA) or Envelope-Tracking Power amplifier (ET-PA), have been attracting much attention because they can maintain high efficiency in large back-off region [1]-[6]. The dynamic power supply voltage techniques cause strong nonlinearity compared to a conventional power amplifier, hence a memoryless Digital Predistortion (DPD) technique is indispensable for these efficiency enhancement techniques. However, the performance of the memoryless DPD is degraded due to the frequency response of the envelope amplifier in the dynamic power supply voltage techniques [7]-[9]. In this paper, we clarify the degradation mechanisms of the memoryless DPD for the EER-PA due to the frequency response of the envelope amplifier based on the results of two-tone tests, and propose an analytical model for improving the performance of the memoryless DPD developed for the EER-PA. In addition, a prototype EER-PA is developed and we demonstrate that the residual distortion of the developed EER-PA with conventional memoryless DPD algorithm is compensated by the new algorithm based on the proposed analytical model. In the two-tone test, third-order intermodulation distortion (IMD3) with a tone spacing from 100 kHz to 4 MHz is improvement by up to 25 dB by the memoryless DPD algorithm based on the proposed model. Measured adjacent channel leakage power ratio (ACPR) of the developed EER-PA is improved from -22.5 dBc to -42.5 dBc in the OFDM signal test with 1.08 MHz bandwidth.

The applications of dynamic content updates for a group of users, for example weather reports and news broadcast, have been shown to benefit significantly from Delay Tolerant Networks (DTNs) communication mechanisms. In this paper, we study the performance of dynamic content updates over DTNs by focusing on the latest content distribution, which is an important factor of the system energy consumption and content update efficiency. By characterizing the content generating process and content sharing process, we obtain an explicit expression for the latest content distribution, and prove it theoretically. Moreover, through simulations based on two synthetical mobility models and a real-world scenario, we demonstrate the accuracy and correctness of the theoretically obtained result.

In this paper, we propose an interleaving updating framework of disparity and confidence map (IUFDCM) for stereo matching to eliminate the redundant and interfere information from unreliable pixels. Compared with other propagation algorithms using matching cost as messages, IUFDCM updates the disparity map and the confidence map in an interleaving manner instead. Based on the Confidence-based Support Window (CSW), disparity map is updated adaptively to alleviate the effect of input parameters. The reassignment for unreliable pixels with larger probability keeps ground truth depending on reliable messages. Consequently, the confidence map is updated according to the previous disparity map and the left-right consistency. The top ranks on Middlebury benchmark corresponding to different error thresholds demonstrate that our algorithm is competitive with the best stereo matching algorithms at present.

In this letter, we apply dynamic software updating to long-lived applications on the DDS middleware while minimizing service interruption and satisfying Quality of Service (QoS) requirements. We dynamically updated applications which run on a commercial DDS implementation to demonstrate the applicability of our approach to dynamic updating. The results show that our update system does not impose an undue performance overhead–all patches could be injected in less than 350 ms and the maximum CPU usage is less than 17%. In addition, the overhead on application throughput due to dynamic updates ranged from 0 to at most 8% and the deadline QoS of the application was satisfied while updating.

We propose a novel re-ranking method for content-based medical image retrieval based on the idea of pseudo-relevance feedback (PRF). Since the highest ranked images in original retrieval results are not always relevant, a naive PRF based re-ranking approach is not capable of producing a satisfactory result. We employ a two-step approach to address this issue. In step 1, a Pearson's correlation coefficient based similarity update method is used to re-rank the high ranked images. In step 2, after estimating a relevance probability for each of the highest ranked images, a fuzzy SVM ensemble based approach is adopted to re-rank the images. The experiments demonstrate that the proposed method outperforms two other re-ranking methods.

This paper presents a new non-interactive string-commitment scheme that achieves universally composable security. Security is proven under the decisional composite residuosity (DCR) assumption (or the decisional Diffie-Hellman (DDH) assumption) in the common reference string (CRS) model. The universal composability (UC) is a very strong security notion. If cryptographic protocols are proven secure in the UC framework, then they remain secure even if they are composed with arbitrary protocols and polynomially many copies of the protocols are run concurrently. Many UC commitment schemes in the CRS model have been proposed, but they are either interactive commitment or bit-commitment (not string-commitment) schemes. We note, however, that although our scheme is the first non-interactive UC string-commitment scheme, a CRS is not reusable. We use an extension of all-but-one trapdoor functions (ABO-TDFs) proposed by Peikert and Waters at STOC 2008 as an essential building block. Our main idea is to extend (original deterministic) ABO-TDFs to probabilistic ones by using the homomorphic properties of their function indices. The function indices of ABO-TDFs consist of ciphertexts of homomorphic encryption schemes (such as ElGamal, and Damgåd-Jurik encryption). Therefore we can re-randomize the output of ABO-TDFs by re-randomization of ciphertexts. This is a new application of ABO-TDFs.

LaMacchia, Lauter and Mityagin [19] proposed a novel security definition for authenticate key exchange (AKE) that gives an adversary the power to obtain ephemeral information regarding a target test session. To demonstrate feasibility of secure protocols in the new definition, henceforth called eCK, the authors described a protocol called NAXOS. NAXOS combines an ephemeral private key x with a static private key a to generate an ephemeral public key X (more precisely in what we call the NAXOS' approach X = gH(x,a)). Thus no one is able to query the discrete logarithm of X without knowing both the ephemeral and static private keys. This idea is crucial in the security argument to guard against leaked ephemeral secrets belonging to the test session. Another important assumption is the gap assumption that allows the protocol to remain secure even in the presence of malicious insiders. Both ideas have been successfully used in creating various protocols secure in the eCK model. In this paper, we construct two eCK-secure protocols without the above mentioned ideas. KFU1 is secure under the GDH assumption without using the NAXOS' approach. KFU2 builds upon KFU1 and drops the gap requirement, thus it is secure under the CDH assumption. Efficiency and security of the proposed protocols are comparable to the well-known HMQV [15] protocol. Furthermore, unlike HMQV and NAXOS the use of the random oracle in KFU1 and KFU2 is restricted to the key derivation function making them more suitable for practical applications.

This paper presents a new non-interactive multi-trapdoor commitment scheme from the standard RSA assumption. Multi-trapdoor commitment is a stronger variant of trapdoor commitment. Its notion was introduced by Gennaro at CRYPTO 2004. Multi-trapdoor commitment schemes are very useful because we can convert a non-interactive multi-trapdoor commitment scheme into a non-interactive and reusable non-malleable commitment scheme by using one-time signature and transform any proof of knowledge into a concurrently non-malleable one (this can be used as concurrently secure identification). Gennaro gave concrete constructions of multi-trapdoor commitment, but its security relies on stronger assumptions, such as the strong RSA assumption and the q-strong Diffie-Hellman assumption as opposed to our construction based on the standard RSA assumption. As a corollary of our results, we constructed a non-interactive and reusable non-malleable commitment scheme from the standard RSA assumption. Our scheme is based on the Hohenberger-Waters (weak) signature scheme presented at CRYPTO 2009. Several non-interactive and reusable non-malleable commitment schemes (in the common reference string model) have been proposed, but they all rely on stronger assumptions (such as the strong RSA assumption). Thus, we give the first construction of a non-interactive and reusable non-malleable commitment scheme from the standard RSA assumption.

A novel Template Updating system for fingerprint verification systems used in mobile applications is introduced in the paper. Based on the proposed method, the system performance is improved much more than the original one. Not only the FRR (False Reject Rate) but also the small overlap problem caused by the very narrow sensor on the mobile phone are solved. Based on the template updating system, templates are replaced with matched inputs towards a target structure which can expand the coverage of templates with large displacement and rotation. By using the test database, the system performance shows the FRR can be reduced by 79% in comparison with the one without template updating procedure. This system was adopted in practical mobile phones in the commercial market in 2009.

Object tracking is a major technique in image processing and computer vision. Tracking speed will directly determine the quality of applications. This paper presents a parallel implementation for a recently proposed scale- and rotation-invariant on-line object tracking system. The algorithm is based on NVIDIA's Graphics Processing Units (GPU) using Compute Unified Device Architecture (CUDA), following the model of single instruction multiple threads. Specifically, we analyze the original algorithm and propose the GPU-based parallel design. Emphasis is placed on exploiting the data parallelism and memory usage. In addition, we apply optimization technique to maximize the utilization of NVIDIA's GPU and reduce the data transfer time. Experimental results show that our GPGPU-based method running on a GTX480 graphics card could achieve up to 12X speed-up compared with the efficiency equivalence on an Intel E8400 3.0 GHz CPU, including I/O time.

In spectrum sensing, if the primary user (PU) signal and the channel noise both follow Gaussian distribution and neither of their probability distribution functions (PDFs) are known, the traditional approaches based on entropy or Likelihood Ratio Test (LRT) etc., become infeasible. To address this problem, we propose a spectrum sensing method that exploits the similarity of PDFs of two time-adjacent detected data sets with cross entropy, while accounting for achieving the detection performance of LRT which is Neyman-Pearson optimal in detecting the primary user. We show that the detection performance of the proposed method asymptotically approximates that of LRT in detecting the PU. The simulation results confirm our analysis.

Aassuming that the depolarization-induced noise generated in the dual-polarized channel is AWGN and spreads uniformly over the whole channel, we derive an effective XPD formula that can be used to estimate the depolarization effects for both partially and completely overlapped channels.

In this letter, we derive a probability density function (PDF) for a modulus of product of two complex-valued Gaussian random variables. The PDF is expressed using Modified Bessel Functions, and the probability distribution is named Gaussian Product Modulus Distribution. Some examples of expectation calculation are provided.

The panorama image obtained by image stitching can have visible artifacts due to the limitation of alignment accuracy and defects of the optical systems. Moreover, conventional image stitching algorithms cannot be directly applied to a real-time video stitching due to its complexity and waving artifacts. In this paper, we propose a real-time content-aware stitching algorithm which not only finds a seam by using path searching based on the greedy method, but also adaptively updates the seam by detecting objects moving toward the seam. Experimental results show that the proposed algorithm can successfully produce stitched video sequences without the waving and ghost artifacts commonly found in conventional stitching algorithms.

Thin films of a divinyl derivative of tetraphenyldiaminobiphenyl DvTPD were prepared by vapor deposition followed by annealing. After annealing at 200°C for 1 h, the film became practically insoluble to organic solvents due to polymerization. Electrical characteristics of the films were measured by current-voltage measurement, time-of-flight measurement, and dielectric measurement. It was found that the hole mobility of DvTPD decreases when the film is polymerized. As a consequence of the decrease of hole mobility, carrier balance in the emissive layer of an organic light emitting diode (OLED) was improved, leading to a higher quantum efficiency and a pure emission spectrum. The dielectric measurement also confirmed the high thermal stability of the polymerized film.