The proposed algorithm has already been applied to a practical sorting system, and the experimental results reveal that it is robust when applied to low quality paper currency. Within each subzone, the defect feature is calculated to estimate the level of contamination. The paper currency image is divided into several overlapping subzones. The defect feature extracted from edge intensity differential is sensitive to the odd edge-information, and is robust to the gray (or edge ) intensity change. To ensure accurate correlation with the subjective feelings of human beings, an edge intensity differential of two images is then constructed from the edge information extracted by the Kirsch operator.
#Wechat windows dev infrastructure database cache registration#
An area-based image registration algorithm is used to overlay the sensed and referenced paper currency images. In this paper, an edge-based algorithm is proposed to detect the scratches and cracks appearing frequently on paper currency. The experiment shows that only 1000 groups of data are adequate to generate the reasoning rules, and the execution time of FTEA increases exponentially with the expansion of determination factors, so, an open computing system only need to select 5 to 6 factors or 1000 sample data for implementation.ĭefect detection is an essential step in paper currency sorting. The FTEA utilizes the rule-based machine learning method and has the advantage of self-learning to get reasoning rules from large amount of input data. In this paper, a fuzzy trust-level evaluating algorithm (FTEA) based on fuzzy logic is put forward to generate the evaluation rules and integrated overall trust-level. And the trust-level of resources changes dynamically with the users' assessments and the service quality that resources provide. In this framework, the trust-level of users changes dynamically according to their behaviors and some other common inspect parameters. In addition, the adaptive scheme can also be used in PI, REM, AVQ, and PD schemes and offers the possibility of optimizing these AQM schemes.įor the purpose of developing a usable trust relationship between the resource consumers (users) and the resource providers (hosts) in an open computing environment, a dynamic trust evaluation framework based on machine learning is proposed. Verified by using NS-2 simulations under a variety of network and traffic situations, the adaptive PIP can achieve faster convergence speed and smaller queue oscillation than PIP, PI, ARED and SPI(self-configuring PI, which is an improved algorithm of PI). Based on adaptive single-neuron PID controller, an adaptive PIP AQM scheme is developed using square error of queue length as performance criteria to consolidate the advantages of single neuron and PIP controller. Among them, PIP is the fusion of PI controller and position feedback compensation and shows better performance under most network conditions, but its parameters can not change with the environments. Several AQM schemes have been proposed to provide low delay and low loss service in best-effort networks in recent studies, such as RED, PI, REM, AVQ, PD, SMVS and PIP. Active queue management (AQM) is an effective method to improve the performance of end-to-end congestion control.
0 kommentar(er)
