Abstract:This paper presents a new texture spectrum descriptor for content-based image retrieval (CBIR). The relation between wavelet transition and texture spectrum method is discussed. This paper proposes the concept of texture pattern equivalent according to texture visual nature to educe more rational texture spectrum descriptor. The proposed texture patterns depict the variational pattern of pixel gray scales in a domain to descript images texture feature. It represents image content in the form of texture spectrum histogram. The symmetry invariant and rotation robusticity are discussed. Comparison with Gabor features using general image dataset indicate that the proposed texture descriptor takes lower extracting time and provides higher image retrieval accuracy.

Abstract:The “registration pattern” between two fingerprints is the optimal registration of each part of one fingerprint with respect to the other fingerprint. Registration patterns generated from imposter’s matching attempts are different from those patterns from genuine matching attempts, although they may share some similarities in the aspect of minutiae. This paper presents an algorithm that utilizes minutiae, associate ridges and orientation fields to determine the registration pattern between two fingerprints and their similarity. The proposed matching scheme has two stages. An offline training stage derives a genuine registration pattern base from a set of genuine matching attempts. Then, an online matching stage registers the two fingerprints and determines the registration pattern. Only if the pattern makes a genuine one, a further fine matching is conducted. The genuine registration pattern base is derived using a set of fingerprints extracted from the NIST Special Database 24. Experimental results on the second FVC2002 database demonstrate the performance of the proposed algorithm.

Abstract:Text-To-Visual speech (TTVS) synthesis by computer can increase the speech intelligibility and make the human-computer interaction interfaces more friendly. This paper describes a Chinese text-to-visual speech synthesis system based on data-driven (sample based) approach, which is realized by short video segments concatenation. An effective method to construct two visual confusion trees for Chinese initials and finals is developed. A co-articulation model based on visual distance and hardness factor is proposed, which can be used in the recording corpus sentence selection in analysis phase and the unit selection in synthesis phase. The obvious difference between bound ary images of the concatenation video segments is smoothed by image morphing technique. By combining with the acoustic Text-To-Speech (TTS) synthesis, a Chinese text-to-visual speech synthesis system is realized.

Abstract:While using continuous time neural network described by the E.Oja. learning rule (Oja-N) for computing real symmetrical matrix eigenvalues and eigenvectors, the initial vector must be on Rn unit hyper-sphere surface, otherwise, the network may produce limit-time overflow. In order to get over this defect, a new neural network (lyNN) algorithm is proposed. By using the analytic solution of the differential equation of lyNN, the following results are received: If initial vector belongs to a space corresponding to certain eigenvector, the lyNN equilibrium vector will converge in this space; If initial vector does not fall into the space corresponding to any eigenvector, the equilibrium vector will belong to the space spanned by eigenvectors corresponding to the maximum eigenvalue. The initial vector maximum space for the lyNN equilibrium vector will fall into space spanned by eigenvectors corresponding to any eigenvalue received. If the initial vector is perpendicular to a known eigenvector, so is the equilibrium vector. The equilibrium vector is on the hyper-sphere surface decided by the initial vector. By using the above results, a method for computing real symmetric matrix eigenvalues and eigenvectors using lyNN is proposed, the validity of this algorithm is exhibited by two examples, indicating that this algorithm does not bring about limit-time overflow. But for Oja-N, if the initial vector is outside the unit hyper-sphere and the matrix is negatively determinant, the neural network will consequentially produce limit-time overflow. Compared with other algorithms based on optimization, lyNN can be realized directly and its computing weight is lighter.

Abstract:A hybrid algorithm based on attention model (HAAM) is proposed to speed up the training of back-propagation neural networks and improve the performances. The algorithm combines the genetic algorithm with the BP algorithm based on magnified error signal. The key to this algorithm lies in the partition of the BP training process into many chips with each chip trained by the BP algorithm. The chips in the same iteration are optimized by the GA operators, and those in different iterations constitute the whole training. Therefore, the HAAM obtains the ability of searching the global optimum solution relying on these operations, and it is easy to be parallelly processed. The simulation experiments show that this algorithm can effectively avoid failure training caused by randomizing the initial weights and thresholds, and solve the slow convergence problem resulted from the Flat-Spots when the error signal becomes too small. Moreover, this algorithm improves the generalization of BP network by improving the training precision instead of adding hidden neurons.

Abstract:Supervised classification based on error-correcting output codes (ECOC) is a new research direction to improve the generalization of classifiers, yet there is no general method to construct ECOC for any number of classes. This paper analyzes the properties of ECOC and presents a search coding method which corresponds to codewords with integers and gets a satisfied output code through searching an integer range in sequence. It then describes the supervised classification technique based on the search coding method. By applying the search coding method to na?ve-Bayes algorithm and BP neural networks, experimental results show that the method is an effective and general coding method to construct error-correcting output codes.

Abstract:Vast computation is a great disadvantage of the existing graph cuts based vision algorithms. Lack of adaptability is another issue. An improved global optimal algorithm for dense disparity mapping using graph cuts is presented in this paper. First, adapted occlusion penalty and smoothness penalty are defined based on the intrinsic relation between the disparity changes and the discontinuities in an image. The graph cuts based algorithm is employed to get an optimial dense disparity mapping with occlusions. Secondly, according to the complexity analysis of graph cut algorithms, an operation named restricted α-expansion operation is defined to control the vertexes generation during graph constructing based on the result of normalized correlation algorithm. It is a great help to reduce the vertexes and edges in the constructed graph, thus the computing is speeded up. The experimental results show performance of the proposed algorithm is improved and it will take a shorter time to compute an accuracy dense disparity mapping.

Abstract:In this paper an efficient approach via wavelet for digital watermark in DEM(digital elevation mode) data is developed, which effectly protects the copyright of DEM data and avoids the unauthoritative user. A technique based on lifting scheme is presented to construct the compactly supported wavelets whose coefficients are composed of a free variable. When t=1, the integer wavelets based on lifting scheme only use integral addition and shift, so it’s fast and easily realized via hardware. A method is presented to build the wavelet coefficient set which can embed watermark information. The bit is inserted in the high activity texture regions with the maximum strength of Just Noticeable Distortion (JND) tolerance of Human Visual System (HVS). Keeping the terrain figure and hypsography, the digital watermark is robust. A hash one-way is constructed by the Rabin method, and the digital watermark arithmetic can be public.

Abstract:Network wide broadcasting is a basic operation in mobile ad hoc networks. Counter-Based broadcasting scheme is well suited for highly mobile networks because of its simplicity. The paper presents several improvements on the counter-based broadcasting scheme and proposes an enhanced counter-based broadcasting scheme, CBB+. Analyses and simulations have shown that CBB+ operates more efficiently than the original counter-based scheme. It has higher reachability, fewer rebroadcasts and lower delay.

Abstract:Due to the difficulties in deployment of IP multicast, the overlay multicast networks (OMN) are being increasingly recognized as a viable alternative to construct a general multicast service infrastructure. To settle the heterogeneity of bandwidth required by real-time multimedia application, the multicast routing problem in heterogeneous OMN is discussed. A new network model of heterogeneous OMN is described by extending the degree-constrained model. Adopting the strategy of layered bandwidth allocation, a heuristic routing algorithm, called layered compact tree (LCT) is proposed to build a minimum delay radius multicast tree in heterogeneous OMN. The properties of LCT are also proved and analyzed in theory. The simulation results show that LCT could reduce both hops and network resource usage of the tree effectively, and keep a little increase in delay radius while the allocated bandwidth decreases.

Abstract:Current assured forwarding (AF) service in differentiated services (DiffServ) networks can provide stable guarantees in throughput, but is lacking of efficient schemes in ensuring queuing delay and loss ratio. By analyzing the steady state operating point of RIO, this paper proposes two active queue management algorithms with adaptive control policy, namely ARIO-D and ARIO-L. These two algorithms can provide differentiated performance in, respectively, queuing delay and loss ratio, in addition to throughput guarantee. By deploying ARIO-D and ARIO-L, AF service can provide quantitative guarantees for multimedia traffic with multiple QoS metrics.

Abstract:Traditional Cluster Area Network(cLAN)’s evaluation model takes only latency, bandwidth, routing, congestion, network topology and some related aspects into consideration. Are these factors ENOUGH to describe the real applications’ communication behavior or predict its performance on cLAN? In the large quantity of NAS Parallel Benchmarks’ tests(version 2.4) on a modern supercomputer——DeepComp 1800, which is of LINUX Cluster architecture, it is found that the real performance of cLAN could be greatly affected by a special communication pattern(LU pattern). Further investigation reveals that the cLAN’s capacity of dealing with LU mode is independent of the known performance factors such as latency, bandwidth and so on. So it is necessary to take some new considerations on cLAN’s evaluation model and add one new factor to reflect the abnormal phenomenon. The new model also provides some challenges in parallel algorithm design and application performance improvement on the LINUX Cluster.

Abstract:Gnutella application layer protocol simply uses flooding algorithm to route peer’s querying, which is just implemented on application layer and doesn’t use down-layer’s information routing of Internet. So it has poor scalability and low efficiency. This paper uses the model of “small world and power law” as the theoretical foundation, and in the light of the users’ requirement and a shared organization layer and region of logic manage-unit, a new distributed peer-to-peer network model of RLP2P(region-layer P2P) with multi-layer structure based on region is proposed, and its prototype system is implemented. A new optimized Multilayer Light-Gossip route strategy is implemented. This paper studies the stress and stretch with the model which has two intuitive metrics of goodness defined to evaluate the quality of the data paths. At the same time, the idea is put forward about the tradeoffs between the stress and stretch. Simulation results about RLP2P’s protocol show that it could effectively solve the above problems, and the larger the network size is, the more obvious the superiority of its comprehensive performance is. So the model is reasonable and effective.

Abstract:Program behavior control is an active detection mechanism. The research of program behavior contro mainly focuses on four aspects: audit data selection, behavior description, the establishment of normal behavior and behavior matching. This paper investigates the event sequence model and proposes the use of LSM(Linux security modules) as an alternative data source to system calls. Based on the data quality analysis and execution results from real systems, the efficiency of the LSM data source is verified from both theoretical and practical points of view Results show that, because of its more refined granularity and its better security relevance, LSM data source is more suitable for the audit events used in event sequence models.

Abstract:In wireless IP networks, it is necessary to adopt header compression schemes for the reduction of the protocol header size in order to save the scarce wireless bandwidth resource and make real-time services economically feasible and physically realizable. However, all the existing header compression schemes have not taken wireless channel state into consideration when designed. For a better usage over wireless links, a new channel state based robust header compression algorithm (CSB-ROHC) for real-time streams in wireless IP networks such as 3G platforms is proposed and analyzed in this paper. Through adjusting the dimension of VSW in W-LSB encoding (a key parameter in header compressor) with the accurate estimation of wireless channel state, this new algorithm can achieve a good balance of compression ratio and error-resistant robustness for its adaptive usage over wireless links. The authors present simulation results that demonstrate the effectiveness of the CSB-ROHC algorithm over wireless links.

Abstract:In Mobile IP, MN detects its movement according to subnet prefix and router advertisement interval which forms a uniform distribution in the region [min,max]. Therefore, the size of this region has a direct impact on movement detection delay. However, this region is configured statically in Mobile IP, which is inconvenient and inflexible. In this paper, this region is adjusted by using the link bandwidth, and a movement detection delay formula is given. In addition, the proposed method is modeled and analyzed. All the results show that the dynamic region, compared with the fixed region, could reduce the unnecessary movement detection delay drastically.

Abstract:In this paper, the authors use SPI calculus to analyze the demand loading code protocol of active network. The security flaw of being replay attack in this protocol is found. Model of active network is different from the tradition network. Since the model of active network is compute-storage-forward, and the model of tradition network is storage-forward. The replay attack would make the unexpected result for active network and letdown the performance and efficiency of active node. In order to prevent the replay attack, the authors amend the protocol and enhance the capacity for preventing replay attack.

Abstract:This paper presents the formal verification of properties of neighbor discovery protocol of IPv6 protocol suite using model checking. The protocol is modeled in MSC, whose use is popular in designing and documenting communication protocols. Linear temporal logic is adopted to specify properties of the protocol. The main result of this paper is an automatic method to extract properties from the MSC linearization directly.

Abstract:concept of specified verifiers is first introduced into the threshold proxy signature, and a (t,n) threshold proxy signature scheme with the specified verifiers is proposed. In the proposed scheme, any t or more proxy signers can sign a message on behalf of an original signer for the specified verifiers, and only the specified verifiers together are able to verify the validity of the proxy signature. In ordinary (t,n) proxy threshold signature schemes, anyone can verify the validity of the proxy signature. In some applications, however, it is required that a proxy signature could be verified only by the specified verifiers together, that is, no a verifier can gain an advantage of the knowledge of the validity of the proxy signature before the other verifiers know its validity (e.g. tenders, bits). Furthermore, the proposed scheme can also revoke the proxy signature right of the proxy signers delegated by the original signer if original signer needs.