The aim of single image super-resolution(SISR) is to find a non-linear mapping from the low to high resolution training image samples and reconstruct a high-resolution(HR) image from a low-resolution(LR) image using t...
详细信息
The aim of single image super-resolution(SISR) is to find a non-linear mapping from the low to high resolution training image samples and reconstruct a high-resolution(HR) image from a low-resolution(LR) image using the prior *** a powerful tool in solving high-level computer vision problems,deep learning can also be used to solve the low-level vision problem,such as *** exploring the architecture of convolutional neural network(CNN),we put forward a robust SISR algorithm based on *** to the state-of-art SISR methods,the proposed SISR algorithm obtains the best SR performance.
Physical and mathematical model of a three-dimensional double physical pendulum being coupled by two universal joints is studied. Upper joint of this pendulum is periodically rotated in its axial direction. Damping fo...
详细信息
According to the energy bottleneck of the wireless sensor network,the LEACH algorithm is improved in this paper, in the LEACH algorithm, it maybe that the cluster head nodes are far from the base station, and cluster ...
详细信息
According to the energy bottleneck of the wireless sensor network,the LEACH algorithm is improved in this paper, in the LEACH algorithm, it maybe that the cluster head nodes are far from the base station, and cluster head node has non-uniform distribution, in this case, nodes soon died after consuming energy, so the focus from three aspects of the optimal number of the cluster head and uniform distribution and many jump communication to improve, the improved algorithm through the simulation tests show that network node energy has been obviously saved, at the same time the lifecycle of the system is also effectively extension. In the application of the wireless sensor, we should always hold the principle of energy-first, therefore the LEACH which is a hierarchical routing protocol can save energy has played a crucial role, but the optimal number of the cluster head and uniform distribution can't be archived because of the change of the location of the cluster head nodes. In addition, when the cluster head far away from center sink node communicate with center sink node, the cluster head will exhaust energy quickly to die, that may affect lifecycle of whole of monitoring network and lead to worse expansibility. For some problems of the LEACH algorithm, the LEACH algorithm is improved in this paper from three aspects of the optimal number of the cluster head and uniform distribution and many jump communication, the improved algorithm through the simulation tests show that network node energy has been obviously saved.
As an effective service mode, online auditing which serves the highest state audit institution is emerging as an important tool for national governance. However, as a crucial part during online auditing workflow, data...
详细信息
ISBN:
(纸本)9781509001545
As an effective service mode, online auditing which serves the highest state audit institution is emerging as an important tool for national governance. However, as a crucial part during online auditing workflow, data stream processing has become one of the most important problems that restrains the development of online auditing. To address this problem, our study are primarily focusing on three research contents: Firstly, a hierarchal online auditing cloud framework applying to data stream processing based on SON (service Overlay Network, SON) is proposed in this paper. Secondly, an efficient virtual resource scheduling and composition strategy called LOSPS (Local Optimal service Path Selection, LOSPS) is also proposed, which adequately taking consideration of demands for underlying resources that satisfies upper application. Thirdly, a dynamic resource selection and composition algorithm in view of multiple attribute decision making and local optimization, which greatly supports the strategy we proposed above, is further presented. At last, various classic resource allocation algorithms including ours are discussed and their performances are compared in a simulated online auditing scenario. Simulation results show that the virtual resource scheduling and composition scheme in LOSPS outperforms the classic service composition algorithms either in static or in dynamic environments. The research findings can greatly improve the reliability and availability of online auditing via adopting the effective virtual resource allocation and composition way we proposed.
It's always difficult to satisfy both strict requirements of dynamic performance and settling performance of high-precision and high-acceleration point-to-point motions on electronic packaging equipment, i.e., Die...
详细信息
It's always difficult to satisfy both strict requirements of dynamic performance and settling performance of high-precision and high-acceleration point-to-point motions on electronic packaging equipment, i.e., Die bonders, Wire bonders, Flip-chip bonders, Wafer bumping machines, and so on. Focusing on this difficulty, Fuzzy logic is promoted in this paper to fine tune feed forward coefficient parameters of motion systems on packaging equipment to satisfy multiple targets of dynamic and settling performance requirements. This intelligent fine tuning approach, with achievable multiple motion targets, provides a therapy to automate and optimize motion system performance on packaging equipment, help on development and mass production of semiconductor packaging equipment, and improve machine intelligence.
The high-obfuscation plagiarism detection in big data environment, such as the paraphrasing and cross-language plagiarism, is often difficult for anti- plagiarism system because the plagiarism skills are becoming more...
详细信息
Recently, various optical networks have been proposed for chip-scale communication on large-scale many-core chips, including optical burst-switching(OBS). The overall throughput of OBS network is dually constrained by...
详细信息
ISBN:
(纸本)9781510821965
Recently, various optical networks have been proposed for chip-scale communication on large-scale many-core chips, including optical burst-switching(OBS). The overall throughput of OBS network is dually constrained by congestion and loss in both the data-plane(i.e., burst contention) and the control-plane(i.e., header packet losses). While previous studies mainly focus on maximizing data-plane throughput, little work has been done on the control-plane congestion problem. In chip-scale OBS, massive fine-grained data bursts, stringent transmission delay budget and moderate network operation frequency have actually constrained the electronic header processing capability, leading to serious network congestion. This paper proposes a new approach based on ingress regulation to overcome the control-plane congestion problem. In our scheme, before being injected, every concurrent packet flow is globally regulated and coordinated so that the aggregated flows do not exceed the header processing capacity of routing nodes, leading to the alleviation of control-plane congestion. We use real application traces to evaluate our method. Simulation results validate that the proposed approach can effectively minimize the control-plane congestion and improve system performance in terms of transmission delay and burst loss rate.
Human Epithelial type 2 (HEp-2) cells are the most common substrates for anti-nuclear antibodies detection. Traditional manual diagnosis heavily depends on the experience of histopathologists, which is time consuming ...
详细信息
ISBN:
(纸本)9781785610448
Human Epithelial type 2 (HEp-2) cells are the most common substrates for anti-nuclear antibodies detection. Traditional manual diagnosis heavily depends on the experience of histopathologists, which is time consuming and subject to subjective mistakes. With the recent progress of digital scanners and dramatic development in computer vision techniques, computer-aided diagnosis has now become achievable. In this paper a novel automatic system is proposed to classify the HEp- 2 cell images into six categories. Along with a set of local gradient based textural descriptors, we introduce a novel objectbased method to decompose the binary image into primitive objects and represent them with a set of morphological features. Random forest is then applied for classification. The advantages of this system are as following: (1) robustness against the changes of intensity and rotation, (2) more discriminative information compared to normal morphological descriptors. We evaluate the proposed approach using the publicly available ICPR 2012 datasets. The experimental results show that the proposed method achieves comparable performance with the state-of-the-art methods.
The hybrid brain computer interface (hBCI) could provide higher information transfer rate than did the classical BCIs. It included more than one brain-computer or human-machine interact paradigms, such as the combinat...
详细信息
ISBN:
(纸本)9781424492701
The hybrid brain computer interface (hBCI) could provide higher information transfer rate than did the classical BCIs. It included more than one brain-computer or human-machine interact paradigms, such as the combination of the P300 and SSVEP paradigms. Research firstly constructed independent subsystems of three different paradigms and tested each of them with online experiments. Then we constructed a serial hybrid BCI system which combined these paradigms to achieve the functions of typing letters, moving and clicking cursor, and switching among them for the purpose of browsing webpages. Five subjects were involved in this study. They all successfully realized these functions in the online tests. The subjects could achieve an accuracy above 90% after training, which met the requirement in operating the system efficiently. The results demonstrated that it was an efficient system capable of robustness, which provided an approach for the clinic application.
This paper is devoted to address the robust tracking problem, where an automatic object detecting and tracking algorithm based on consecutive frames differencing and the improved mean-shift is presented. The moving ta...
详细信息
ISBN:
(纸本)9781941763162
This paper is devoted to address the robust tracking problem, where an automatic object detecting and tracking algorithm based on consecutive frames differencing and the improved mean-shift is presented. The moving target is obtained by differencing the previous two frames in the video sequence. Then, by differencing in the framework of mean-shift tracking, the size of box is updated and the position of target is corrected. It has been shown that the proposed method inherits the advantages of the two algorithms, and also make up for the disadvantages. Experimental results demonstrate that the proposed method reduces system’s sensitivity to target’s moving speed and the color of background, and more importantly shows higher accuracy and more stability. Keywords: Moving object tracking;computer vision;Frame Differencing;Mean-shift algorithm.
暂无评论