An efficient search method for large-vocabulary continuous-speech recognition

Recognizing individuals by their gait is a new biometric methodology, which employs dynamic features derived from tracking gait. Instead of the image processing techniques used in most existing studies, our previous study initialized the work of investigating gait recognition in terms of biomechanics. The experimental results showed that the angles and forces of the lower limb joints were reliable features for recognition of individuals, which can provide us with a considerable amount of information in the field of computer science and thus help in developing a more efficient recognition method, which is also more computationally efficient than current image processing methods. Encouraged by the early results, in this study, we proposed a people recognition method based on plantar pressure patterns, which can be used in a concealed manner. We hoped to prove the feasibility of using foot pressure for individual recognition.Two different plantar pressure parameter measurement schemes are discussed: (1) the characteristic parameters and (2) the pressure values of each sensor cell in each frame. The self-organizing map (SOM) neuron network algorithm was used in both schemes for data classification. In order to improve the recognition rate, a support vector machine (SVM) was used as the data classification algorithm for the all-sensor-values method. High recognition rates were achieved with the second method, i.e., using all the sensor cell values of the foot pressure pattern during walking, regardless of the algorithm used. It is suggested that the foot pressure distribution of gait is a suitable feature for gait recognition. Both SOM and SVM can be feasible classifiers for foot pressure-based features.

Scalable and queryable compressed storage structure for raster data

Forwarding packets based on networking names is essential for network protocols on different layers, where the `names' could be addresses, packet/flow IDs, and content IDs. For long there have been efforts using dynamic and compact data structures for fast and memory-efficient forwarding. In this work, we identify that the recently developed programmable network paradigm has the potential to further reduce the time/memory complexity of forwarding structures by separating the data plane and control plane. This work presents the new designs of network forwarding structures under the programmable network paradigm, applying three typical dynamic and compact data structures: Bloom filters, Cuckoo hashing, and Othello hashing. We conduct careful analyses and experiments in real networks of these forwarding methods for multiple performance metrics, including lookup throughput, memory footprint, construction time, dynamic updates, and lookup errors. The results give rich insights on designing forwarding algorithms with dynamic and compact data structures. In particular, the new designs based on Cuckoo hashing and Othello hashing show significant advantages over the extensively studied Bloom filter based methods, in all situations discussed in this paper.

Human recognition based on gait is a new biometric method that considers both spatial and temporal features which tracks the gait of a human being from a distance. The most widely used techniques for gait recognition at present can be divided into two categories: the model-free method and the model-based method. Both these methods process images of a moving body. Although some effort has been devoted to gait recognition, a high identification rate and low computational cost still need to be achieved. In this study, we investigated gait recognition in terms of biomechanics, which is often overlooked in the fields of computer science and electronics. We present a method for people recognition by using both the three-dimensional parameters for lower limb joints, i.e., kinematic and kinetic parameters. Dynamic data are acquired by tracking marker positions using a motion capture system with force plates. Although a marker-free and cable-free biometric system is more applicable for security purposes, acquiring data by marker tracking is one of the most precise methods currently available. The results obtained from marker tracking can clarify the relationship between the recognition rate and dynamic data of normal walking. In this study, 350 trials were conducted for 10 subjects. For each subject, 20 trials were conducted for training and the remaining 15 were for testing. A self-organizing map (SOM) neural network was employed for data classification. The importance of a specific variable in each dimension for each joint is discussed in detail to investigate the major factors that cause the differences in human gait according to a biomechanical approach. Experimental results showed that gait is a reliable feature for individual identification since a high recognition rate can be achieved by choosing appropriate joints or dynamic parameters in some dimensions of a gait. Kinematic variables in the frontal and transverse planes of the knee or hip joints are recommended due to their higher recognition rates. These findings can be applied to both biomedicine and computer science.

The objective is to develop a data structure that is capable of handling large data volumes and offers support for querying, analysis and validation. Based on earlier results (i.e. the full decomposition of space, the use of a TEN structure and applying Poincare simplicial homology as mathematical foundation) a simplicial complex-based TEN structure is developed. Applying simplicial homology offers full control over orientation of simplexes and enables one to derive substantial parts of the TEN structure, instead of explicitly store the entire network. The described data structure is developed as a DBMS data structure and the usage of views, function based indexes and 3D R-trees result in a compact topological 3D data structure. Theoretical aspects of this approach are described earlier [12, 17,18]. This paper describes both theory and implementation of the approach.

Streaming triangle counting is a critical issue in graph stream mining, with applications in dense subgraph discovery, web mining, anomaly detection, and more. Recent efforts have focused on estimating triangle counts in graph streams, primarily through sampling methods. However, because of limited memory resources for handling high speed streams, traditional sampling methods suffer from reduced sampling rate and thereby performance loss. In this paper, we propose a new compact data structure called uHLL to process edge streams by considering the tradeoff between estimation accuracy and memory efficiency. Furthermore, different from conventional triangle counting algorithms, we solve the estimation of union set cardinality for edge-local triangle count under both centralized and distributed framework, so as to efficiently estimate the global triangle count by a one-pass streaming algorithm. To the best of our knowledge, this is the first implementation of a distributed framework using a compact data structure for streaming triangle counting. We provide theoretical proof of unbiasedness and derive the variance of the union set and global triangle count. We compare our scheme with 11 algorithms, showing that under the same experimental setting, uHLL and distributed uHLL are at least <inline-formula><tex-math notation="LaTeX">$ 2.3$</tex-math></inline-formula> and <inline-formula><tex-math notation="LaTeX">$ 1.7$</tex-math></inline-formula> times more accurate than the state-of-the-art, respectively.

Using Compressed Suffix-Arrays for a compact representation of temporal-graphs

The convex differences tree (CDT) representation of a simple polygon is useful in computer graphics, computer vision, computer aided design and robotics. The root of the tree contains the convex hull of the polygon and there is a child node recursively representing every connectivity component of the set difference between the convex hull and the polygon. We give an O(n log K + K log2 n) time algorithm for constructing the CDT, where n is the number of polygon vertices and K is the number of nodes in the CDT. The algorithm is adaptive to a complexity measure defined on its output while still being worst case efficient. For simply shaped polygons, where K is a constant, the algorithm is linear. In the worst case K = O(n) and the complexity is O(n log2 n). We also give an O(n log n) algorithm which is an application of the recently introduced compact interval tree data structure.

VDB[Museth 2013] is a compact data structure and toolset developed at DreamWorks Animation for high-resolution volumetric effects typically encountered in movie production. Since its open source release in 2012, as OpenVDB1, it has been adopted by major third-party renders and VFX tools, including Houdini by SideFX, RenderMan by Pixar, Arnold by Solid Angle, and RealFlow by Next Limit. Thus, it should come as no surprise that it is highly desirable to have efficient algorithms for ray-marching of sparse VDB volumes. The core problem is therefore, how to best utilize the hierarchical data structure of VDB for applications that require efficient ray-marching. As we will demonstrate, the solution is a novel hierarchical digital differential analyzer.

: Problem Statement: Pecan weevil is one of the most destructive pests of Oklahoma. The scope of this study is to develop a recognition system that can serve in a wireless imaging network for monitoring pecan weevils. Approach: The recognition methods used in this study are based on template matching. Five recognition methods were implemented: Normalized cross-correlation, Fourier descriptors, Zernike moments, String matching and Regional properties. The training set consisted of 205 pecan weevils and the testing set included 30 randomly selected pecan weevils and 74 other insects which typically exist in pecan habitat. Results: It was found that Region-based methods were better in representing and recognizing biological objects such as insects. Different recognition rates were obtained at different order of Zernike moments. The optimum result among the tested orders of Zernike moments was found to be at the order 3. The results also showed that using different number of Fourier descriptors may not significantly increase the recognition rate of this method. Conclusion: The most robust and reliable recognition rate was achieved when the Zernike moments and Region properties recognition methods were used in a combination. A positive match from either of these two independent tests would yield reliable results. Therefore, 100% recognition could be achieved by adopting the proposed algorithm. The processing time for such recognition is 0.44 sec.

The classic string indexing problem is to preprocess a string S into a compact data structure that supports efficient pattern matching queries. Typical queries include existential queries (decide if the pattern occurs in S), reporting queries (return all positions where the pattern occurs), and counting queries (return the number of occurrences of the pattern). In this paper we consider a variant of string indexing, where the goal is to compactly represent the string such that given two patterns \(P_1\) and \(P_2\) and a gap range \({[}\alpha , \beta ]\) we can quickly find the consecutive occurrences of \(P_1\) and \(P_2\) with distance in \({[}\alpha , \beta ]\), i.e., pairs of subsequent occurrences with distance within the range. We present data structures that use linear space and query time \({\widetilde{O}}(|P_1|+|P_2|+n^{2/3})\) for existence and counting and \({\widetilde{O}}(|P_1|+|P_2|+n^{2/3}\hbox {occ}^{1/3})\) for reporting. We complement this with a conditional lower bound based on the set intersection problem showing that any solution using \({\widetilde{O}}(n)\) space must use \({\widetilde{\Omega }}(|P_1| + |P_2| + \sqrt{n})\) query time. To obtain our results we develop new techniques and ideas of independent interest including a new suffix tree decomposition and hardness of a variant of the set intersection problem.

We present a new multithreaded implementation for the computationally demanding direct volume rendering (DVR) of volumetric data sets on desktop multicore processors using ray casting. The new implementation achieves interactive rendering of very large volumes, even on high resolution screens. Our implementation is based on a new algorithm that combines an object‐order traversal of the volumetric data followed by a focused ray casting. Using a very compact data structure, our method starts with a quick association of data subcubes with fine‐grain screen tiles appearing along the viewing direction in front‐to‐back order. The next stage uses very limited ray casting on the generated sets of subcubes while skipping empty or transparent space and applying early ray termination in an effective way. Our multithreaded implementation makes use of new dynamic techniques to ensure effective memory management and load balancing. Our software enables a user to interactively explore large data sets through DVR while arbitrarily specifying a 2D transfer function. We test our system on a wide variety of well‐known volumetric data sets on a two‐processor Clovertown platform, each consisting of a Quad‐Core 1.86 GHz Intel Xeon Processor. Our experimental tests demonstrate DVR at interactive rates for the largest data sets that can fit in the main memory on our platform. These tests also indicate a high degree of scalability, excellent load balancing, and efficient memory management across the data sets used. Copyright © 2009 John Wiley &amp; Sons, Ltd.

The Metric Suffix Array (MSA) is a variant of the classical suffix array designed to support similarity queries over multimedia data. It is a permutation-base index that stores lists of objects ranked by their similarity to a set of reference objects. In this paper, we present two compact alternatives for representing the Metric Suffix Array using less space. The first uses variable-length integer encoding to store the difference between consecutive values in MSA buckets, which are decoded during the queries with a small overhead in time. The second encodes the values in the concatenated ranking list using a Wavelet tree, allowing to obtain MSA values without explicitly computing them. All compact data structures are constructed directly from the input data without the need to compute the complete MSA beforehand, thereby enabling the indexing of much larger data volumes. Experimental results with high-dimensional features extracted from 64 million images showed that our compact alternatives require less than 25% of the original method’s total memory while increasing the query execution time by a small constant factor.

Efficient pebbling for list traversal synopses with application to program rollback

Many services used in large scale web applications should be able to tolerate faults without impacting their performance. State machine replication is a well-known approach to implementing fault-tolerant services, providing high availability and strong consistency. To boost the performance of state machine replication, recent proposals have introduced parallel execution of commands. In parallel state machine replication, incoming commands may or may not depend on other commands that are waiting for execution. Although dependent commands must be processed in the same relative order at every replica to avoid inconsistencies, independent commands can be executed in parallel and benefit from multi-core architectures. Since many application workloads are mostly composed of independent commands, these parallel models promise high throughput without sacrificing strong consistency. The efficient execution of commands in such environments, however, requires effective scheduling strategies. Existing approaches rely on dependency tracking based on pairwise comparison between commands, which introduces scheduling contention. In this paper, we propose a new and highly efficient scheduler for parallel state machine replication. Our scheduler considers batches of commands, instead of commands individually. Moreover, each batch of commands is augmented with a compact data structure that encodes commands information needed to the dependency analysis. We show, by means of experimental evaluation, that our technique outperforms schedulers for parallel state machine replication by a fairly large margin.