Robust object tracking is a crucial enabler of many computer vision applications. An embedded object tracker must, in addition to respecting real-time requirements, robustly update the position of the tracked object i...
详细信息
ISBN:
(纸本)9781450354875
Robust object tracking is a crucial enabler of many computer vision applications. An embedded object tracker must, in addition to respecting real-time requirements, robustly update the position of the tracked object in the captured images and be able to re-identify the object when it reappears after disappearing for a number of frames. To obtain such a robustness, modern algorithms such as the Tracking Learning Detection (TLD) combine classical tracking techniques with learning capabilities. this combination makes it possible to track an object's displacement, learn its new appearances when it changes orientation and re-identify it once it reappears. the strong performances of these techniques are obtained using costly computation that challenges hardware capabilities in the context of embedded architectures. this paper proposes a strategy for porting the TLD algorithm on an Field-Programmable Gate Array (FPGA). Based on parallel image resizing and selective object patch rejection, the proposed dataflow-oriented strategy demonstrates how the structure of the algorithm can be adapted to be more streaming-oriented and fit the structure of an FPGA. the results prove a tracking performance of up to 160 fps under 50Mhz for VGA images while keeping nearly the tracking robustness of the reference algorithm implementation.
UML (Unified Modeling Language) is designed for all stages of software development, but it lacks precise semantic information. MDE (Model-driven engineering) takes the model as the primary software product and its mai...
详细信息
ISBN:
(纸本)9789811064425;9789811064418
UML (Unified Modeling Language) is designed for all stages of software development, but it lacks precise semantic information. MDE (Model-driven engineering) takes the model as the primary software product and its main research direction is modeling and model transformation. We propose to explore the entity abstraction scenario where no existing classes fit as the representative entity for other classes through the data, information and knowledge recreation of existing classes and relationships withthe introduction of knowledge graphs. In this paper, we proposed a knowledge graph to enhance model design between models.
Analogical proportions, i.e., statements of the form a is to b as c is to d, are supposed to obey 3 axioms expressing reflexivity, symmetry, and stability under central permutation. these axioms are not enough to dete...
详细信息
ISBN:
(纸本)9783319615813;9783319615806
Analogical proportions, i.e., statements of the form a is to b as c is to d, are supposed to obey 3 axioms expressing reflexivity, symmetry, and stability under central permutation. these axioms are not enough to determine a single Boolean model, if a minimality condition is not added. After an algebraic discussion of this minimal model and of related expressions, another justification of this model is given in terms of Kolmogorov complexity. It is shown that the 6 Boolean patterns that make an analogical proportion true have a minimal complexity with respect to an expression reflecting the intended meaning of the proportion.
Modification of single and complex nickel-palladium samples by laser processing in the femtosecond time domain was studied. the samples were processed by focused Ti:Sapphire laser beam (Clark CPA-2101) with 775 nm las...
详细信息
Modification of single and complex nickel-palladium samples by laser processing in the femtosecond time domain was studied. the samples were processed by focused Ti:Sapphire laser beam (Clark CPA-2101) with 775 nm laser wavelength, 2 kHz repetition rate, 200 fs pulse duration. the laser-induced morphological modifications have shown dependence on the applied fluences and number of laser pulses. the formed surface nanostructures on the single NiPd/Si and multilayer 5x(Ni/Pd)/Si systems are compared with individual Ni and Pd thin films. the results show an increase in surface roughness, formation of parallel periodic surface structures, appearance of hydrodynamic features and ablation of surface material. At low number of pulses (less than 10 pulses) and low pulse energies range (not over 1.7 mu J), the two types of laser-induced periodic surface structure (LIPSS) can be observed: low and high spatial frequency LIPSS (HSFL and LSFL). For all samples, the measured LSFL periods were 720 nm for the ripples created solely on thin film surfaces during the single pulse action. In the case of the multi-pulse irradiation, the periodicities of created LSFLs on the all investigated thin films have shown tendency to reduction with increasing of pulse energies. (C) 2016 Elsevier B.V. All rights reserved.
In parallel computing, a valid graph coloring yields a lock-free processing of the colored tasks, data points, etc., without expensive synchronization mechanisms. However, coloring is not free and the overhead can be ...
详细信息
In parallel computing, a valid graph coloring yields a lock-free processing of the colored tasks, data points, etc., without expensive synchronization mechanisms. However, coloring is not free and the overhead can be significant. In particular, for the bipartite-graph partial coloring (BGPC) and distance-2 graph coloring (D2GC) problems, which have various use-cases within the scientific computing and numerical optimization domains, the coloring overhead can be in the order of minutes with a single thread for many real-life *** this work, we propose parallelalgorithms for bipartite-graph partial coloring on shared-memory architectures. Compared to the existing shared-memory BGPC algorithms, the proposed ones employ greedier and more optimistic techniques that yield a better parallel coloring performance. In particular, on 16 cores, the proposed algorithms are more than 4x faster than their counterparts in the ColPack library which is, to the best of our knowledge, the only publicly-available coloring library for multicore architectures. In addition to BGPC, the proposed techniques are employed to devise parallel distance-2 graph coloring algorithms and similar performance improvements have been observed. Finally, we propose two costless balancing heuristics for BGPC that can reduce the skewness and imbalance on the cardinality of color sets (almost) for free. the heuristics can also be used for the D2GC problem and in general, they will probably yield a better color-based parallelization performance especially on many-core architectures.
this paper presents a system that uses machine learning to recognize military vehicles in social media images. To do so, the system draws on recent advances in applying deep neural networks to computer vision tasks, w...
详细信息
ISBN:
(纸本)9781509067275
this paper presents a system that uses machine learning to recognize military vehicles in social media images. To do so, the system draws on recent advances in applying deep neural networks to computer vision tasks, while also making extensive use of openly available libraries, models and data. Training a vehicle recognition system over three classes, the paper reports on two experiments that use different architectures and strategies to overcome the challenges of working with limited training data: data augmentation and transfer learning. the results show that transfer learning outperforms data augmentation, achieving an average accuracy of 95.18% using 10-fold cross-validation, while also generalizing well on a separate testing set consisting of social media content.
the paper presents a study on the efficiency measures of the Ant Colony Communities (ACC). the ACC is an approach to parallelize the Ant Colony Optimization algorithm (ACO). An ACC is made up of a Community Server tha...
详细信息
ISBN:
(纸本)9783319439822;9783319439815
the paper presents a study on the efficiency measures of the Ant Colony Communities (ACC). the ACC is an approach to parallelize the Ant Colony Optimization algorithm (ACO). An ACC is made up of a Community Server that coordinates the work of a set Ant Colony clients. Each client implements a classical ACO algorithm. the individual colonies work in an asynchronous manner processing data sent by server and sending back the obtained results. there are many possible locations for the clients: the same computer as the server, computers of a local or wide area network. the paper presents a detailed description of concept the ACC and reports the study of the efficiency of the such Communities. the efficiency is measured by their power (the amount of data processed in a given period of time) and scalability-the efficiency of adding colony clients on the Community. the paper contains also the taxonomy of parallel implementations of the Ant Colony.
A commercial Graphics processing Unit (GPU) is used to build a fast Level 0 (L0) trigger system tested parasitically withthe TDAQ (Trigger and Data Acquisition systems) of the NA62 experiment at CERN. In particular, ...
详细信息
A commercial Graphics processing Unit (GPU) is used to build a fast Level 0 (L0) trigger system tested parasitically withthe TDAQ (Trigger and Data Acquisition systems) of the NA62 experiment at CERN. In particular, the parallel computing power of the GPU is exploited to perform real-time fitting in the Ring Imaging CHerenkov (RICH) detector. Direct GPU communication using a FPGA-based board has been used to reduce the data transmission latency. the performance of the system for multi-ring reconstrunction obtained during the NA62 physics run will be presented.
As the volume of data and technical complexity of large-scale analysis increases, many domain experts can no longer be seated in the data exploration and analysis workflow. What is desired is a computational powerful ...
详细信息
ISBN:
(纸本)9781450355490
As the volume of data and technical complexity of large-scale analysis increases, many domain experts can no longer be seated in the data exploration and analysis workflow. What is desired is a computational powerful but still familiar analysis interface for domain experts to fully participate in the analysis workflow by just focusing on individual datasets, leaving the large-scale computation to the system. Towards this goal, we present VisRden, a research prototype that combines user friendly visual programming and scalable computing backend for large-scale 3D reconstruction in carious lesion research. VisRden uses R as the analysis language, making a set of core functions available to the users by hiding the computational complexity behind a visual interface, and allowing advanced users to provide custom R scripts and variables to be fully embedded into the final analysis script. Using R as the analysis language allows cariologists to continue explore data and propose new analysis methods in the way they are already familiar with. VisRden conquers large-scale image processing and 3D reconstruction in a MapReduce-like framework using R and SGE (Sun Grid Engine) array jobs. Image-based operations and result aggregation are scheduled as array jobs in a parallel means to accelerate the knowledge discovery process. All these combine to provide a new analytics workflow for performing similar large-scale analysis loops that need expert users to closely supervise, provide feedback, and refine the subtasks.
Cognitive capabilities can enhance a business process by offering automated analytics-based recommendations on key decisions by applying machine learning techniques. Yet the organizational adoption of such advanced ca...
详细信息
暂无评论