We propose a parallel precomputation method for real-time model predictive control. The key idea is to use predicted input values produced by model predictive control to solve an optimal control problem in advance. It is well known that control systems are not suitable for multi- or many-core processors because feedback-loop control systems are inherently based on sequential operations. However, since the proposed method does not rely on conventional thread-/data-level parallelism, it can be easily applied to such control systems without changing the algorithm in applications. A practical evaluation using three real-world model predictive control system simulation programs demonstrates drastic performance improvement without degrading control quality offered by the proposed method.

As energy efficiency has become a major design constraint or objective, heterogeneous manycore architectures have emerged as mainstream target platforms not only in server systems but also in embedded systems. Manycore accelerators such as GPUs are getting also popular in embedded domains, as well as the heterogeneous CPU cores. However, as the number of cores in an embedded GPU is far less than that of a server GPU, it is important to utilize both heterogeneous multi-core CPUs and GPUs to achieve the desired throughput with the minimal energy consumption. In this paper, we present a case study of mapping LBP-based face detection onto a recent CPU-GPU heterogeneous embedded platform, which exploits both task parallelism and data parallelism to achieve maximal energy efficiency with a real-time constraint. We first present the parallelization technique of each task for the GPU execution, then we propose performance and energy models for both task-parallel and data-parallel executions on heterogeneous processors, which are used in design space exploration for the optimal mapping. The design space is huge since not only processor heterogeneity such as CPU-GPU and big.LITTLE, but also various data partitioning ratios for the data-parallel execution on these heterogeneous processors are considered. In our case study of LBP face detection on Exynos 5422, the estimation error of the proposed performance and energy models were on average -2.19% and -3.67% respectively. By systematically finding the optimal mappings with the proposed models, we could achieve 28.6% less energy consumption compared to the manual mapping, while still meeting the real-time constraint.

The response times of solid state drives (SSDs) have decreased dramatically due to the growing use of non-volatile memory express (NVMe) devices. Such devices have response times of less than 100 micro seconds on average. The response times of all-flash-array systems have also decreased dramatically through the use of NVMe SSDs. However, there are applications, particularly virtual desktop infrastructure and in-memory database systems, that require storage systems with even shorter response times. Their workloads tend to contain many input-output (IO) concentrations, which are aggregations of IO accesses. They target narrow regions of the storage volume and can continue for up to an hour. These narrow regions occupy a few percent of the logical unit number capacity, are the target of most IO accesses, and appear at unpredictable logical block addresses. To drastically reduce the response times for such workloads, we developed an automated tiered storage system called “automated tiered storage with fast memory and slow flash storage” (ATSMF) in which the data in targeted regions are migrated between storage devices depending on the predicted remaining duration of the concentration. The assumed environment is a server with non-volatile memory and directly attached SSDs, with the user applications executed on the server as this reduces the average response time. Our system predicts the effect of migration by using the previously monitored values of the increase in response time during migration and the change in response time after migration. These values are consistent for each type of workload if the system is built using both non-volatile memory and SSDs. In particular, the system predicts the remaining duration of an IO concentration, calculates the expected response-time increase during migration and the expected response-time decrease after migration, and migrates the data in the targeted regions if the sum of response-time decrease after migration exceeds the sum of response-time increase during migration. Experimental results indicate that ATSMF is at least 20% faster than flash storage only and that its memory access ratio is more than 50%.

In a multiprocessor system, processors are connected based on various types of network topologies. A network topology is usually represented by a graph. Let G be a graph and u, v be any two distinct vertices of G. We say that G is pancyclic if G has a cycle C of every length l(C) satisfying 3≤l(C)≤|V(G)|, where |V(G)| denotes the total number of vertices in G. Moreover, G is panpositionably pancyclic from r if for any integer m satisfying $r \leq m \leq \frac{|V(G)|}{2}$, G has a cycle C containing u and v such that d_C(u,v)=m and 2m≤l(C)≤|V(G)|, where d_C(u,v) denotes the distance of u and v in C. In this paper, we investigate the panpositionable pancyclicity problem with respect to the n-dimensional locally twisted cube LTQ_n, which is a popular topology derived from the hypercube. Let D(LTQ_n) denote the diameter of LTQ_n. We show that for n≥4 and for any integer m satisfying $D(LTQ_n) + 2 \leq m \leq \frac{|V(LTQ_n)|}{2}$, there exists a cycle C of LTQ_n such that d_C(u,v)=m, where (i) 2m+1≤l(C)≤|V(LTQ_n)| if m=D(LTQ_n)+2 and n is odd, and (ii) 2m≤l(C)≤|V(LTQ_n)| otherwise. This improves on the recent result that u and v can be positioned with a given distance on C only under the condition that l(C)=|V(LTQ_n)|. In parallel and distributed computing, if cycles of different lengths can be embedded, we can adjust the number of simulated processors and increase the flexibility of demand. This paper demonstrates that in LTQ_n, the cycle embedding containing any two distinct vertices with a feasible distance is extremely flexible.

The order/degree problem consists of finding the smallest diameter graph for a given order and degree. Such a graph is beneficial for designing low-latency networks with high performance for massively parallel computers. The average shortest path length (ASPL) of a graph has an influence on latency. In this paper, we propose a novel order adjustment approach. In the proposed approach, we search for Cayley graphs of the given degree that are close to the given order. We then adjust the order of the best Cayley graph to meet the given order. For some order and degree pairs, we explain how to derive the smallest known graphs from the Graph Golf 2016 and 2017 competitions.

Generalized recursive circulant graphs (GRCGs for short) are a generalization of recursive circulant graphs and provide a new type of topology for interconnection networks. A graph of n vertices is said to be s-pancyclic for some $3\leqslant s\leqslant n$ if it contains cycles of every length t for $s\leqslant t\leqslant n$. The pancyclicity of recursive circulant graphs was investigated by Araki and Shibata (Inf. Process. Lett. vol.81, no.4, pp.187-190, 2002). In this paper, we are concerned with the s-pancyclicity of GRCGs.

Datacenter growth in traffic and scale is driving innovations in constructing tightly-coupled facilities with low-latency communication for different specific applications. A famous custom design is rackscale (RS) computing by gathering key server resource components into different resource pools. Such a resource-pooling implementation requires a new software stack to manage resource discovery, resource allocation and data communication. The reconfiguration of interconnection networks on their components is potentially needed to support the above demand in RS. In this context as an evolution of the original RS architecture the inter-rackscale (IRS) architecture, which disaggregates hardware components into different racks according to their own areas, has been proposed. The heart of IRS is to use a limited number of free-space optics (FSO) channels for wireless connections between different resource racks, via which selected pairs of racks can communicate directly and thus resource-pooling requirements are met without additional software management. In this study we evaluate the influences of FSO links on IRS networks. Evaluation results show that FSO links reduce average communication hop count for user jobs, which is close to the best possible value of 2 hops and thus provides comparable benchmark performance to that of the counterpart RS architecture. In addition, if four FSO terminals per rack are allowed, the CPU/SSD (GPU) interconnection latency is reduced by 25.99% over Fat-tree and by 67.14% over 2-D Torus. We also present the advantage of an FSO-equipped IRS system in average turnaround time of dispatched jobs for given sets of benchmark workloads.

Video analytics is usually time-consuming as it not only requires video decoding as a first step but also usually applies complex computer vision and machine learning algorithms to the decoded frame. To achieve high efficiency in video analytics with ever increasing frame size, many researches have been conducted for distributed video processing using Hadoop. However, most approaches focused on processing multiple video files on multiple nodes. Such approaches require a number of video files to achieve any speedup, and could easily result in load imbalance when the size of video files is reasonably long since a video file itself is processed sequentially. In contrast, we propose a distributed video decoding method with an extended FFmpeg and VideoRecordReader, by which a single large video file can be processed in parallel across multiple nodes in Hadoop. The experimental results show that a case study of face detection and SURF system achieve 40.6 times and 29.1 times of speedups respectively on a four-node cluster with 12 mappers in each node, showing good scalability.

The Infrastructure as a Service (IaaS) Clouds are emerging as a promising platform for the execution of resource demanding and computation intensive workflow applications. Scheduling the execution of scientific applications expressed as workflows on IaaS Clouds involves many uncertainties due to the variable and unpredictable performance of Cloud resources. These uncertainties are modeled by probability distribution functions in past researches or totally ignored in some cases. In this paper, we propose a novel robust deadline constrained workflow scheduling algorithm which handles the uncertainties in scheduling workflows in the IaaS Cloud environment. Our proposal is a static scheduling algorithm aimed at addressing the uncertainties related to: the estimation of task execution times; and, the delay in provisioning computational Cloud resources. The workflow scheduling problem was considered as a cost-optimized, deadline-constrained optimization problem. Our uncertainty handling strategy was based on the consideration of knowledge of the interval of uncertainty, which we used to modeling the execution times rather than using a known probability distribution function or precise estimations which are known to be very sensitive to variations. Experimental evaluations using CloudSim with synthetic workflows of various sizes show that our proposal is robust to fluctuations in estimates of task runtimes and is able to produce high quality schedules that have deadline guarantees with minimal penalty cost trade-off depending on the length of the interval of uncertainty. Scheduling solutions for varying degrees of uncertainty resisted against deadline violations at runtime as against the static IC-PCP algorithm which could not guarantee deadline constraints in the face of uncertainty.

Data replication in cloud storage systems brings a lot of benefits, such as fault tolerance, data availability, data locality and load balancing both from reliability and performance perspectives. However, each time a datanode fails, data blocks stored on the failed datanode must be restored to maintain replication level. This may be a large burden for the system in which resources are highly utilized with users' application workloads. Although there have been many proposals for replication, the approach of re-replication has not been properly addressed yet. In this paper, we present a deferred re-replication algorithm to dynamically shift the re-replication workload based on current resource utilization status of the system. As workload pattern varies depending on the time of the day, simulation results from synthetic workload demonstrate a large opportunity for minimizing impacts on users' application workloads with the simple algorithm that adjusts re-replication based on current resource utilization. Our approach can reduce performance impacts on users' application workloads while ensuring the same reliability level as default HDFS can provide.

A parallel phrase matching (PM) engine for dictionary compression is presented. Hardware based parallel chaining hash can eliminate erroneous PM results raised by hash collision; while newly-designed storage architecture holding PM results solved the data dependency issue; Thus, the average compression speed is increased by 53%.

This paper intends to reduce duration times in typical collective communications. We introduce logical addressing system apart from the physical one and, by rearranging the logical node addresses properly, we intend to reduce communication overheads so that ideal communication is performed. One of the key issues is rearrangement of the logical addressing system. We introduce genetic algorithm (GA) as meta-heuristic solution as well as the random search strategy. Our GA-based method achieves at most 2.50 times speedup in three-traffic-pattern cases.

Recent studies utilize multiple kernel learning to deal with incomplete-data problem. In this study, we introduce new methods that do not only complete multiple incomplete kernel matrices simultaneously, but also allow control of the flexibility of the model by parameterizing the model matrix. By imposing restrictions on the model covariance, overfitting of the data is avoided. A limitation of kernel matrix estimations done via optimization of an objective function is that the positive definiteness of the result is not guaranteed. In view of this limitation, our proposed methods employ the LogDet divergence, which ensures the positive definiteness of the resulting inferred kernel matrix. We empirically show that our proposed restricted covariance models, employed with LogDet divergence, yield significant improvements in the generalization performance of previous completion methods.

Dai-Yuan (DY) conjugate gradient method is an effective method for solving large-scale unconstrained optimization problems. In this paper, a new DY method, possessing a spectral conjugate parameter β_k, is presented. An attractive property of the proposed method is that the search direction generated at each iteration is descent, which is independent of the line search. Global convergence of the proposed method is also established when strong Wolfe conditions are employed. Finally, comparison experiments on impulse noise removal are reported to demonstrate the effectiveness of the proposed method.

Kernel updates are a part of daily life in contemporary computer systems. They usually require an OS reboot that involves restarting not only the kernel but also all of the running applications, causing downtime that can disrupt software services. This downtime issue has been tackled by numerous approaches. Although dynamic translation of the running kernel image, which is a representative approach, can conduct kernel updates at runtime, its applicability is inherently limited. This paper describes Dwarf, which shortens downtime during kernel updates and covers more types of updates. Dwarf launches the newer kernel in the background on the same physical machine and forces the kernel to inherit the running states of the older kernel. We implemented a prototype of Dwarf on Xen 4.5.2, Linux 2.6.39, Linux 3.18.35, and Linux 4.1.6. Also, we conducted experiments using six applications, such as Apache, MySQL, and memcached, and the results demonstrate that Dwarf's downtime is 1.8 seconds in the shortest case and up to 10× shorter than that of the normal OS reboot.

The security of a software program critically depends on the prevention of vulnerabilities in the source code; however, conventional computer programs lack the ability to identify vulnerable code in another program. Our research was aimed at developing a technique capable of generating substitution code for the detection of buffer overflow vulnerability in C/C++ programs. The technique automatically verifies and sanitizes code instrumentation by comparing the result of each candidate variable with that expected from the input data. Our results showed that statements containing buffer overflow vulnerabilities could be detected and prevented by using a substitution variable and by sanitizing code vulnerabilities based on the size of the variables. Thus, faults can be detected prior to execution of the statement, preventing malicious access. Our approach is particularly useful for enhancing software security monitoring, and for designing retrofitting techniques in applications.

Graphics processing units (GPUs) have become an attractive platform for general-purpose computing (GPGPU) in various domains. Making GPUs a time-multiplexing resource is a key to consolidating GPGPU applications (apps) in multi-tenant cloud platforms. However, advanced GPGPU apps pose a new challenge for consolidation. Such highly functional GPGPU apps, referred to as GPU eaters, can easily monopolize a shared GPU and starve collocated GPGPU apps. This paper presents GLoop, which is a software runtime that enables us to consolidate GPGPU apps including GPU eaters. GLoop offers an event-driven programming model, which allows GLoop-based apps to inherit the GPU eaters' high functionality while proportionally scheduling them on a shared GPU in an isolated manner. We implemented a prototype of GLoop and ported eight GPU eaters on it. The experimental results demonstrate that our prototype successfully schedules the consolidated GPGPU apps on the basis of its scheduling policy and isolates resources among them.

In multiprocessors, memory models are introduced to describe the executions of programs among processors. Relaxed memory models, which relax the order of executions, are used in the most of the modern processors, such as ARM and POWER. Due to a relaxed memory model could change the program semantics, the executions of the programs might not be the same as our expectation that should preserve the program correctness. In addition to relaxed memory models, the way to execute an instruction is described by an instruction semantics, which varies among processor architectures. Dealing with instruction semantics among a variety of assembly programs is a challenge for program verification. Thus, this paper proposes a way to verify a variety of assembly programs that are executed under a relaxed memory model. The variety of assembly programs can be abstracted as the way to execute the programs by introducing an operation structure. Besides, there are existing frameworks for modeling relaxed memory models, which can realize program executions to be verified with a program property. Our work adopts an SMT solver to automatically reveal the program executions under a memory model and verify whether the executions violate the program property or not. If there is any execution from the solver, the program correctness is not preserved under the relaxed memory model. To verify programs, an experimental tool was developed to encode the given programs for a memory model into a first-order formula that violates the program correctness. The tool adopts a modeling framework to encode the programs into a formula for the SMT solver. The solver then automatically finds a valuation that satisfies the formula. In our experiments, two encoding methods were implemented based on two modeling frameworks. The valuations resulted by the solver can be considered as the bugs occurring in the original programs.

In many applications, tables are distributively stored in different data sources, but the frequency of updates on each data source is different. Some techniques have been proposed to effectively express the temporal orders between different values, and the most current, i.e. up-to-date, value of a given data item can be easily picked up according to the temporal orders. However, the currency of the data items in the same table may be different. That is, when a user asks for a table D, it cannot be ensured that all the most current values of the data items in D are stored in a single table. Since different data sources may have overlaps, we can construct a conjunctive query on multiple tables to get all the required current values. In this paper, we formalize the conjunctive query as currency preserving query, and study how to generate the minimized currency preserving query to reduce the cost of visiting different data sources. First, a graph model is proposed to represent the distributed tables and their relationships. Based on the model, we prove that a currency preserving query is equivalent to a terminal tree in the graph, and give an algorithm to generate a query from a terminal tree. After that, we study the problem of finding minimized currency preserving query. The problem is proved to be NP-hard, and some heuristics strategies are provided to solve the problem. Finally, we conduct experiments on both synthetic and real data sets to verify the effectiveness and efficiency of the proposed techniques.

In this paper, we propose a highly efficient mobile visual search algorithm. For descriptor extraction process, we propose a low complexity feature detection which utilizes the detected local key points of the coarse octaves to guide the scale space construction and feature detection in the fine octave. The Gaussian and Laplacian operations are skipped for the unimportant area, and thus the computing time is saved. Besides, feature selection is placed before orientation computing to further reduce the complexity of feature detection by pre-discarding some unimportant local points. For the image retrieval process, we design a high-performance reranking method, which merges both the global descriptor matching score and the local descriptor similarity score (LDSS). In the calculating of LDSS, the tf-idf weighted histogram matching is performed to integrate the statistical information of the database. The results show that the proposed highly efficient approach achieves comparable performance with the state-of-the-art for mobile visual search, while the descriptor extraction complexity is largely reduced.

Malware has become a growing threat as malware writers have learned that signature-based detectors can be easily evaded by packing the malware. Packing is a major challenge to malware analysis. The generic unpacking approach is the major solution to the threat of packed malware, and it is based on the intrinsic nature of the execution of packed executables. That is, the original code should be extracted in memory and get executed at run-time. The existing generic unpacking approaches need a simulated environment to monitor the executing of the packed executables. Unfortunately, the simulated environment is easily detected by the environment-sensitive packers. It makes the existing generic unpacking approaches easily evaded by the packer. In this paper, we propose a novel unpacking approach, BareUnpack, to monitor the execution of the packed executables on the bare-metal operating system, and then extracts the hidden code of the executable. BareUnpack does not need any simulated environment (debugger, emulator or VM), and it works on the bare-metal operating system directly. Our experimental results show that BareUnpack can resist the environment-sensitive packers, and improve the unpacking effectiveness, which outperforms other existing unpacking approaches.

Hidden Singer is a television program in Korea. In the show, the original singer and four imitating singers sing a song in hiding behind a screen. The audience and TV viewers attempt to guess who the original singer is by listening to the singing voices. Usually, there are few correct answers from the audience, because the imitators are well trained and highly skilled. We propose a computerized system for distinguishing the original singer from the imitating singers. During the training phase, the system learns only the original singer's song because it is the one the audience has heard before. During the testing phase, the songs of five candidates are provided to the system and the system then determines the original singer. The system uses a 1-class authentication method, in which only a subject model is made. The subject model is used for measuring similarities between the candidate songs. In this problem, unlike other existing studies that require artist identification, we cannot utilize multi-class classifiers and supervised learning because songs of the imitators and the labels are not provided during the training phase. Therefore, we evaluate the performances of several 1-class learning algorithms to choose which one is more efficient in distinguishing an original singer from among highly skilled imitators. The experiment results show that the proposed system using the autoencoder performs better (63.33%) than other 1-class learning algorithms: Gaussian mixture model (GMM) (50%) and one class support vector machines (OCSVM) (26.67%). We also conduct a human contest to compare the performance of the proposed system with human perception. The accuracy of the proposed system is found to be better (63.33%) than the average accuracy of human perception (33.48%).

Recent researches on mobile robots show that convolutional neural network (CNN) has achieved impressive performance in visual place recognition especially for large-scale dynamic environment. However, CNN leads to the large space of image representation that cannot meet the real-time demand for robot navigation. Aiming at this problem, we evaluate the feature effectiveness of feature maps obtained from the layer of CNN by variance and propose a novel method that reserve salient feature maps and make adaptive binarization for them. Experimental results demonstrate the effectiveness and efficiency of our method. Compared with state of the art methods for visual place recognition, our method not only has no significant loss in precision, but also greatly reduces the space of image representation.

Latent variable models for relational data enable us to extract the co-cluster structure underlying observed relational data. The Infinite Relational Model (IRM) is a well-known relational model for discovering co-cluster structures with an unknown number of clusters. The IRM and several related models commonly assume that the link probability between two objects depends only on their cluster assignment. However, relational models based on this assumption often lead us to extract many non-informative and unexpected clusters. This is because the cluster structures underlying real-world relationships are often blurred by biases of individual objects. To overcome this problem, we propose a multi-layered framework, which extracts a clear de-blurred co-cluster structure in the presence of object biases. Then, we propose the Multi-Layered Infinite Relational Model (MLIRM) which is a special instance of the proposed framework incorporating the IRM as a co-clustering model. Furthermore, we reveal that some relational models can be regarded as special cases of the MLIRM. We derive an efficient collapsed Gibbs sampler to perform posterior inference for the MLIRM. Experiments conducted using real-world datasets have confirmed that the proposed model successfully extracts clear and interpretable cluster structures from real-world relational data.

For voice-enabled car navigation systems that use a multi-purpose cloud speech recognition service (cloud ASR), utterance classification that is robust against speech recognition errors is needed to realize a user-friendly voice interface. The purpose of this study is to improve the accuracy of utterance classification for voice-enabled car navigation systems when inputs to a classifier are error-prone speech recognition results obtained from a cloud ASR. The role of utterance classification is to predict which car navigation function a user wants to execute from a spontaneous utterance. A cloud ASR causes speech recognition errors due to the noises that occur when traveling in a car, and the errors degrade the accuracy of utterance classification. There are many methods for reducing the number of speech recognition errors by modifying the inside of a speech recognizer. However, application developers cannot apply these methods to cloud ASRs because they cannot customize the ASRs. In this paper, we propose a system for improving the accuracy of utterance classification by modifying both speech-signal inputs to a cloud ASR and recognized-sentence outputs from an ASR. First, our system performs speech enhancement on a user's utterance and then sends both enhanced and non-enhanced speech signals to a cloud ASR. Speech recognition results from both speech signals are merged to reduce the number of recognition errors. Second, to reduce that of utterance classification errors, we propose a data augmentation method, which we call “optimal doping,” where not only accurate transcriptions but also error-prone recognized sentences are added to training data. An evaluation with real user utterances spoken to car navigation products showed that our system reduces the number of utterance classification errors by 54% from a baseline condition. Finally, we propose a semi-automatic upgrading approach for classifiers to benefit from the improved performance of cloud ASRs.

Depth image based rendering (DIBR), which is utilized to render virtual views with a color image and the corresponding depth map, is one of the key procedures in the 2D to 3D conversion process. However, some troubling problems, such as depth edge misalignment, disocclusion occurrences and cracks at resampling, still exist in current DIBR systems. To solve these problems, in this paper, we present a robust depth image based rendering scheme for stereoscopic view synthesis. The cores of the proposed scheme are two depth map filters which share a common domain transform based filtering framework. As a first step, a filter of this framework is carried out to realize texture-depth boundary alignments and directional disocclusion reduction smoothing simultaneously. Then after depth map 3D warping, another adaptive filter is used on the warped depth maps with delivered scene gradient structures to further diminish the remaining cracks and noises. Finally, with the optimized depth map of the virtual view, backward texture warping is adopted to retrieve the final texture virtual view. The proposed scheme enables to yield visually satisfactory results for high quality 2D to 3D conversion. Experimental results demonstrate the excellent performances of the proposed approach.

This paper proposes a block-permutation-based encryption (BPBE) scheme for the encryption-then-compression (ETC) system that enhances the color scrambling. A BPBE image can be obtained through four processes, positional scrambling, block rotation/flip, negative-positive transformation, and color component shuffling, after dividing the original image into multiple blocks. The proposed scheme scrambles the R, G, and B components independently in positional scrambling, block rotation/flip, and negative-positive transformation, by assigning different keys to each color component. The conventional scheme considers the compression efficiency using JPEG and JPEG 2000, which need a color conversion before the compression process by default. Therefore, the conventional scheme scrambles the color components identically in each process. In contrast, the proposed scheme takes into account the RGB-based compression, such as JPEG-LS, and thus can increase the extent of the scrambling. The resilience against jigsaw puzzle solver (JPS) can consequently be increased owing to the wider color distribution of the BPBE image. Additionally, the key space for resilience against brute-force attacks has also been expanded exponentially. Furthermore, the proposed scheme can maintain the JPEG-LS compression efficiency compared to the conventional scheme. We confirm the effectiveness of the proposed scheme by experiments and analyses.

In this paper, a perceptual distortion based rate-distortion optimized video coding scheme for High Efficiency Video Coding (HEVC) is proposed. Structural Similarity Index (SSIM) in transform domain, which is known as distortion metric to better reflect human's perception, is derived for the perceptual distortion model to be applied for hierarchical coding block structure of HEVC. A SSIM-quantization model is proposed using the properties of DCT and high resolution quantization assumption. The SSIM model is obtained as the sum of SSIM in each Coding Unit (CU) depth of HEVC, which precisely predict SSIM values for the hierarchical quadtree structure of CU in HEVC. The rate model is derived from the entropy, based on Laplacian distributions of transform residual coefficients and is jointly combined with the SSIM-based distortion model for rate-distortion optimization in an HEVC video codec and can be compliantly applied to HEVC. The experimental results demonstrate that the proposed method achieves 8.1% and 4.0% average bit rate reductions in rate-SSIM performance for low-delay and random access configurations respectively, outperforming other existing methods. The proposed method provides better visual quality than the conventional mean square error (MSE)-based RDO coding scheme.

We propose a feature design method for a mobile visual search based on binary features and a bag-of-visual words framework. In mobile visual search, detection error and quantization error are unavoidable due to viewpoint changes and cause performance degradation. Typical approaches to visual search extract features from a single view of reference images, though such features are insufficient to manage detection and quantization errors. In this paper, we extract features from multiview synthetic images. These features are selected according to our novel reliability measure which enables robust recognition against various viewpoint changes. We regard feature selection as a maximum coverage problem. That is, we find a finite set of features maximizing an objective function under certain constraints. As this problem is NP-hard and thus computationally infeasible, we explore approximate solutions based on a greedy algorithm. For this purpose, we propose novel constraint functions which are designed to be consistent with the match conditions in the visual search method. Experiments show that the proposed method improves retrieval accuracy by 12.7 percentage points without increasing the database size or changing the search procedure. In other words, the proposed method enables more accurate search without adversely affecting the database size, computational cost, and memory requirement.

This paper presents an approach for developing an algorithm for automatic license plate recognition system (ALPR) on complex scenes. A plate-style classification method is also proposed in this paper to address the inherent challenges for ALPR in a system that uses multiple plate-styles (e.g., different fonts, multiple plate lay-out, variations in character sequences) which is the case in the current Philippine license plate system. Methods are proposed for each ALPR module: plate detection, character segmentation, and character recognition. K-nearest neighbor (KNN) is used as a classifier for character recognition together with a proposed confidence scoring to rate the decision made by the classifier. A small dataset of Philippine license plates but with relevant features of complex scenarios for ALPR is prepared. Using the proposed system on the prepared dataset, the performance of the system is evaluated on different categories of complex scenes. The proposed algorithm structure shows promising results and yielded an overall accuracy higher than the existing ALPR systems on the dataset consisting mostly of complex scenes.

In real-time 3D ball tracking of sports analysis in computer vision technology, complex algorithms which assure the accuracy could be time-consuming. Particle filter based algorithm has a large potential to accelerate since the algorithm between particles has the chance to be paralleled in heterogeneous CPU-GPU platform. Still, with the target multi-view 3D ball tracking algorithm, challenges exist: 1) serial flowchart for each step in the algorithm; 2) repeated processing for multiple views' processing; 3) the low degree of parallelism in reweight and resampling steps for sequential processing. On the CPU-GPU platform, this paper proposes the double stream system flow, the view priority based threads allocation, and the binary search oriented reweight. Double stream system flow assigns tasks which there is no data dependency exists into different streams for each frame processing to achieve parallelism in system structure level. View priority based threads allocation manipulates threads in multi-view observation task. Threads number is view number multiplied by particles number, and with view priority assigning, which could help both memory accessing and computing achieving parallelism. Binary search oriented reweight reduces the time complexity by avoiding to generate cumulative distribution function and uses an unordered array to implement a binary search. The experiment is based on videos which record the final game of an official volleyball match (2014 Inter-High School Games of Men's Volleyball held in Tokyo Metropolitan Gymnasium in Aug. 2014) and the test sequences are taken by multiple-view system which is made of 4 cameras locating at the four corners of the court. The success rate achieves 99.23% which is the same as target algorithm while the time consumption has been accelerated from 75.1ms/frame in CPU environment to 3.05ms/frame in the proposed system which is 24.62 times speed up, also, it achieves 2.33 times speedup compared with basic GPU implemented work.

In this paper, we propose a novel method to extrinsically calibrate a camera to a 3D reference object that is not directly visible from the camera. We use a human cornea as a spherical mirror and calibrate the extrinsic parameters from the reflections of the reference points. The main contribution of this paper is to present a cornea-reflection-based calibration algorithm with a simple configuration: five reference points on a single plane and one mirror pose. In this paper, we derive a linear equation and obtain a closed-form solution of extrinsic calibration by introducing two ideas. The first is to model the cornea as a virtual sphere, which enables us to estimate the center of the cornea sphere from its projection. The second is to use basis vectors to represent the position of the reference points, which enables us to deal with 3D information of reference points compactly. We demonstrate the performance of the proposed method with qualitative and quantitative evaluations using synthesized and real data.

A tremendous amount of knowledge is present in the ever-growing scientific literature. In order to efficiently grasp such knowledge, various computational tasks are proposed that train machines to read and analyze scientific documents. One of these tasks, Scientific Relation Extraction, aims at automatically capturing scientific semantic relationships among entities in scientific documents. Conventionally, only a limited number of commonly used knowledge bases, such as Wikipedia, are used as a source of background knowledge for relation extraction. In this work, we hypothesize that unannotated scientific papers could also be utilized as a source of external background information for relation extraction. Based on our hypothesis, we propose a model that is capable of extracting background information from unannotated scientific papers. Our experiments on the RANIS corpus [1] prove the effectiveness of the proposed model on relation extraction from scientific articles.

Word boundary ambiguity in word segmentation has long been a fundamental challenge within Thai language processing. The Conditional Random Fields (CRF) model is among the best-known methods to have achieved remarkably accurate segmentation. Nevertheless, current advancements appear to have left the problem of compound words unaccounted for. Compound words lose their meaning or context once segmented. Hence, we introduce a dictionary-based word-merging algorithm, which merges all kinds of compound words. Our evaluation shows that the algorithm can accomplish a high-accuracy of word segmentation, with compound words being preserved. Moreover, it can also restore some incorrectly segmented words. Another problem involving a different word-chunking approach is sentence boundary ambiguity. In tackling the problem, utilizing the part of speech (POS) of a segmented word has been found previously to help boost the accuracy of CRF-based sentence segmentation. However, not all segmented words can be tagged. Thus, we propose a POS-based word-splitting algorithm, which splits words in order to increase POS tags. We found that with more identifiable POS tags, the CRF model performs better in segmenting sentences. To demonstrate the contributions of both methods, we experimented with three of their applications. With the word merging algorithm, we found that intact compound words in the product of topic extraction can help to preserve their intended meanings, offering more precise information for human interpretation. The algorithm, together with the POS-based word-splitting algorithm, can also be used to amend word-level Thai-English translations. In addition, the word-splitting algorithm improves sentence segmentation, thus enhancing text summarization.

Learning semantic representation for translation context is beneficial to statistical machine translation (SMT). Previous efforts have focused on implicitly encoding syntactic and semantic knowledge in translation context by neural networks, which are weak in capturing explicit structural syntax information. In this paper, we propose a new neural network with a tree-based convolutional architecture to explicitly learn structural syntax information in translation context, thus improving translation prediction. Specifically, we first convert parallel sentences with source parse trees into syntax-based linear sequences based on a minimum syntax subtree algorithm, and then define a tree-based convolutional network over the linear sequences to learn syntax-based context representation and translation prediction jointly. To verify the effectiveness, the proposed model is integrated into phrase-based SMT. Experiments on large-scale Chinese-to-English and German-to-English translation tasks show that the proposed approach can achieve a substantial and significant improvement over several baseline systems.

Defects in spacecraft software may result in loss of life and serious economic damage. To avoid such consequences, the software development process incorporates code review activity. A code review conducted by a third-party organization independently of a software development team can effectively identify defects in software. However, such review activity is difficult for third-party reviewers, because they need to understand the entire structure of the code within a limited time and without prior knowledge. In this study, we propose a tool to visualize inter-module dataflow for source code of spacecraft software systems. To evaluate the method, an autonomous rover control program was reviewed using this visualization. While the tool does not decreases the time required for a code review, the reviewers considered the visualization to be effective for reviewing code.

In terms of spatial online aggregation, traditional stand-alone serial methods gradually become limited. Although parallel computing is widely studied nowadays, there scarcely has research conducted on the index-based parallel online aggregation methods, specifically for spatial data. In this letter, a parallel multilevel indexing method is proposed to accelerate spatial online aggregation analyses, which contains two steps. In the first step, a parallel aR tree index is built to accelerate aggregate query locally. In the second step, a multilevel sampling data pyramid structure is built based on the parallel aR tree index, which contribute to the concurrent returned query results with certain confidence degree. Experimental and analytical results verify that the methods are capable of handling billion-scale data.

Finding widespread events in a distributed network is crucial when detecting cyber-attacks or network malfunctions. We propose a new detection scheme for widespread events based on bitmaps that can succinctly record and deliver event information between monitoring agents and a central coordinator. Our proposed scheme reduces communication overhead as well as total number of rounds, and achieves even higher accuracy, compared with the current state of the art.

Crack detection is a vital task to maintain a bridge's health and safety condition. Traditional computer-vision based methods easily suffer from disturbance of noise and clutters for a real bridge inspection. To address this limitation, we propose a two-stage crack detection approach based on Convolutional Neural Networks (CNN) in this letter. A predictor of small receptive field is exploited in the first detection stage, while another predictor of large receptive field is used to refine the detection results in the second stage. Benefiting from data fusion of confidence maps produced by both predictors, our method can predict the probability belongs to cracked areas of each pixel accurately. Experimental results show that the proposed method is superior to an up-to-date method on real concrete surface images.

This letter proposes a novel speech enhancement system based on the ‘L’ shaped triple-microphone. The modified coherence-based algorithm and the first-order differential beamforming are combined to filter the spatial distributed noise. The experimental results reveal that the proposed algorithm achieves significant performance in spatial filtering under different noise scenarios.

We propose a new context-adaptive arithmetic coding (CAAC) scheme for lossless bit rate reduction of parametric stereo (PS) in enhanced aacPlus. Based on the probability analysis of stereo parameters indexes in PS, we propose a stereo band-dependent CAAC scheme for PS. We also propose a new coding structure of the scheme which is simple but effective. The proposed scheme has normal and memory-reduced versions, which are superior to the original and conventional schemes and guarantees significant bit rate reduction of PS. The proposed scheme can be an alternative to the original PS coding scheme at low bit rate, where coding efficiency is very important.

Recently, image forensics community has paid attention to the research on the design of effective algorithms based on deep learning technique. And facts proved that combining the domain knowledge of image forensics and deep learning would achieve more robust and better performance than the traditional schemes. Instead of improving algorithm performance, in this paper, the safety of deep learning based methods in the field of image forensics is taken into account. To the best of our knowledge, this is the first work focusing on this topic. Specifically, we experimentally find that the method using deep learning would fail when adding the slight noise into the images (adversarial images). Furthermore, two kinds of strategies are proposed to enforce security of deep learning-based methods. Firstly, a penalty term to the loss function is added, which is the 2-norm of the gradient of the loss with respect to the input images, and then an novel training method is adopt to train the model by fusing the normal and adversarial images. Experimental results show that the proposed algorithm can achieve good performance even in the case of adversarial images and provide a security consideration for deep learning-based image forensics.

Surface light field advances conventional light field rendering techniques by utilizing geometry information. Using surface light field, real-world objects with complex appearance could be faithfully represented. This capability could play an important role in many VR/AR applications. However, an accurate geometric model is needed for surface light field sampling and processing, which limits its wide usage since many objects of interests are difficult to reconstruct with their usually very complex appearances. We propose a novel two-step optimization framework to reduce the dependency of accurate geometry. The key insight is to treat surface light field sampling as a multi-view multi-texture optimization problem. Our approach can deal with both model inaccuracy and image to model misalignment, making it possible to create high-fidelity surface light field models without using high-precision special hardware.

Many single model methods have been applied to real-time short-term traffic flow prediction. However, since traffic flow data is mixed with a variety of ingredients, the performance of single model is limited. Therefore, we proposed Multi-Long-Short Term Memory Models, which improved traffic flow prediction accuracy comparing with state-of-the-art models.