Non-maximum suppression (NMS) plays a key role in many modern object detectors. It is responsible to remove detection boxes that cover the same object. NMS greedily selects the detection box with maximum score; other detection boxes are suppressed when the degree of overlap between these detection boxes and the selected box exceeds a predefined threshold. Such a strategy easily retain some false positives, and it limits the ability of NMS to perceive nearby objects in cluttered scenes. This paper proposes an effective method combining harmony search algorithm and NMS to alleviate this problem. This method regards the task of NMS as a combination optimization problem. It seeks final detection boxes under the guidance of an objective function. NMS is applied to each harmony to remove imprecise detection boxes, and the remaining boxes are used to calculate the fitness value. The remaining detection boxes in a harmony with highest fitness value are chosen as the final detection results. The standard Pattern Analysis, Statistical Modeling and Computational Learning Visual Object Classes dataset and the Microsoft Common Objects in Context dataset are used in all of the experiments. The proposed method is applied to two popular detection networks, namely Faster Region-based Convolutional Neural Networks and Region-based Fully Convolutional Networks. The experimental results show that the proposed method improves the average precision of these two detection networks. Moreover, the location performance and average recall of these two detectors are also improved. 相似文献
In this work, novel application of evolutionary computational heuristics is presented for parameter identification problem of nonlinear Hammerstein controlled auto regressive auto regressive (NHCARAR) systems through global search competency of backtracking search algorithm (BSA), differential evolution (DE) and genetic algorithms (GAs). The mean squared error metric is used for the fitness function of NHCARAR system based on difference between actual and approximated design variables. Optimization of the cost function is conducted with BSA for NHCARAR model by varying degrees of freedom and noise variances. To verify and validate the worth of the presented scheme, comparative studies are carried out with its counterparts DE and GAs through statistical observations by means of weight deviation factor, root of mean squared error, and Thiel’s inequality coefficient as well as complexity measures. 相似文献
The Fourth Amendment to the U.S. Constitution functions as a shield against excess governmental or police power by prohibiting unreasonable searches and seizures. Since its ratification, legal challenges have tempered this shield by frequently disputing the application of investigative processes and tools, including those that bypass the traditional – and simpler – analysis that focused on physical trespass. But recent technological advancements have prompted novel challenges and have forced the U.S. Supreme Court to adopt a parallel inquiry that evaluates society’s expectations of privacy as an alternate path to invoke the Fourth Amendment’s protections apart from any physical trespass. As revolutionary technology continues to present unique issues, this 200-year-old shield manifests a reflective luster as if polished by years of legal discourse that reveals the priorities of those who would interpret its text. Viewing the Fourth Amendment’s shield as a mirror illustrates not only the thoughts of the drafters that revolved primarily around protecting property interests but also the expectations of modern society with its insistence on promoting privacy. And where the drafters channeled their outrage against the loathsome writs of assistance in colonial times, later Americans continued to denounce the similarly invasive general warrants and attempts by investigators to expand the tools in their arsenal beyond constitutional bounds, especially in the surveillance context. Yet, the problems posed by new technology upon privacy concerns are best resolved by relying on the core principles supporting the Fourth Amendment, previous U.S. Supreme Court precedent, and current societal perspectives regarding privacy as a top priority proven by recently enacted legislation both foreign and domestic.
By applying a similar method to address advancing communication technology and its use as a surveillance tool in Carpenter v. United States, the Court turned this shield-become-mirror upon society to conclude that cell phone location information deserves Fourth Amendment protection because of its untiring comprehensiveness and its uniquely detailed nature. Moreover, nearly every American adult carries a cell phone with them almost all the time, making it possible to create a time-stamped map of any cell-phone-carrying-individual’s movements reaching back years and years. Unfortunately, the Carpenter Court did not extend this crucial protection far enough to protect all cell phone location data, and the unmistakable gap in its holding leaves a potential privacy vulnerability the exploitation of which could cause greater harm than all previously disputed surveillance technology combined because of cell phone usage’s general – near universal – applicability. Allowing cell phone location information to be obtained without probable cause and a proper search warrant not only fails to meet the spirit of the Fourth Amendment, it also begins to tarnish that shield such that it no longer reflects historical or current societal values, reducing its goal of protecting Americans to a hollow incantation of words left to languish as time (and technology) marches on. 相似文献
The unburned H2 can be used to reduce NO emission in conventional TWC (three-way catalyst) for a hydrogen internal combustion engine when it works at equivalence ratio marginally higher than the stoichiometric ratio. To explore the effects and feasibility of this reaction, a Perfectly Stirred Reactor simulation model of TWC has been built with simplified mechanisms. Experiments on a 2.3 L turbocharged hydrogen engine are used to verify the conclusion. It shows that rising initial temperature accelerates the reduction of NO and the maximum reaction rate occurs at 400 °C temperature. The conversion efficiency of NO remains approximately 0 when temperatures below 300 °C. The efficiency reaches a peak value of approximately 98% with 400 °C and declines gradually. The unburned H2 to NO mixing ratio greater than 1.5 in TWC guarantees 100% NO conversion efficiency. The experiments indicate that the NOx concentration decreases from 2056 ppm to 41 ppm at the stoichiometric ratio after the treatment of TWC and NOx reaches 0 ppm with a rich ratio. Results also demonstrate that the suitable reaction temperatures for TWC locate in the range of 400 °C–500 °C. Therefore, if the temperature and the mixing ratio are appropriate, it can achieve zero emissions with NOx reduction by unburned H2 in conventional TWC for a hydrogen engine. 相似文献