In the tree classifier with top-down search, a global decision is made via a series of local decisions. Although this approach gains in classification efficiency, it also gives rise to error accumulation which can be very harmful when the number of classes is very large. To overcome this difficulty, a new tree classifier with the following characteristics is proposed: 1) fuzzy logic search is used to find all ``possible correct classes,' and some similarity measures are used to determine the ``most probable class'; 2) global training is applied to generate extended terminals in order to enhance the recognition rate; 3) both the training and search algorithms have been given a lot of flexibility, to provide tradeoffs between error and rejection rates, and between the recognition rate and speed. A computer simulation of the decision trees for the recognition of 3200 Chinese character categories yielded a very high recognition rate of 99.93 percent and a very high speed of 861 samples/s, when the program was written in a high level language and run on a large multiuser time-sharing computer. 相似文献
This paper discusses how local measurements of positions and surface normals may be used to identify and locate overlapping objects. The objects are modeled as polyhedra (or polygons) having up to six degrees of positional freedom relative to the sensors. The approach operates by examining all hypotheses about pairings between sensed data and object surfaces and efficiently discarding inconsistent ones by using local constraints on: distances between faces, angles between face normals, and angles (relative to the surface normals) of vectors between sensed points. The method described here is an extension of a method for recognition and localization of nonoverlapping parts previously described in [18] and [15]. 相似文献
In this correspondence, a parallel 2-D convolution scheme is presented. The processing structure is a mesh connected array processor consisting of the same number of simple processing elements as the number of pixels in the image. For most windows considered, the number of computation steps required is the same as that of the coefficients of a convolution window. The proposed scheme can be easily extended to convolution windows of arbitrary size and shape. The basic idea of the proposed scheme is to apply the 1-D systolic concept to 2-D convolution on a mesh structure. The computation is carried out along a path called a convolution path in a systolic manner. The efficiency of the scheme is analyzed for windows of various shapes. The ideal convolution path is a Hamiltonian path ending at the center of the window, the length of which is equal to the number of window coefficients. The simple architecture and control strategy make the proposed scheme suitable for VLSI implementation. 相似文献
The response has been studied of eight unfilled elastomers (four natural rubber compounds, epoxidized natural rubber ENR50, butyl rubber, polybutadiene and polyurethane) to erosion by 150 m silica sand at an impact velocity of 48 m sec–1. All were tested at an impact angle of 30°, close to the impact conditions occurring in pipe bends, while two were also tested at 90°. Wide variations in erosion rate were observed between different rubbers. These differences did not correlate systematically with the nature of the base elastomer, glass transition temperature or mechanical properties such as hardness, tensile strength or ultimate tensile elongation, nor with resistance to abrasive wear. Good correlation was, however, found with rebound resilience. A high resilience tended to imply high erosion resistance, The erosion rate was found empirically to be proportional to the quantity (1 - fractional resilience) raised to the power 1.4. Different surface morphologies were found in specimens of high and low resilience after erosion at 30°. Finely spaced transverse ridges formed on high-resilience rubbers, whereas surface features on low-resilience rubbers showed no directionality. Preliminary conclusions are drawn about possible mechanisms of material removal. 相似文献
Conclusions In an MR reactor performance tests of 16 fuel assemblies, with elements having essentially the same structure as standard VVÉR-100 fuel elements, were carried out. Tests of five more fuel assemblies are continuing. Of the 16 assemblies, 13 were studied in a hot laboratory.The tests in the MR, carried out at high loads and with a large number of transition processes, as well as the postreactor studies, indicated that fuel elements of the specified design (with initial helium pressures of 1.96–2.45 MPa) have a high reliability. None of the elements of the fuel assemblies studied malfunctioned due to design defects or faults in their fabrication. During the tests the jackets were subject to a little oxidation and hydrogenation (zirconium-oxide film<3 m thick, hydrogen content less than 0.008% by mass), and their plasticity remained high (the relative elongation at the working temperature remained at the 20% level).Translated from Atomnaya Énergiya, Vol. 62, No. 5, pp. 312–317, May, 1987. 相似文献
Catalysis Letters - The NdM0.25Ni0.75O3 (M?=?Cr, Fe) named NCN and NFN catalysts precursors were synthesized and characterized. The CO2 utilization in the feed was studied in the... 相似文献
Layered perovskite Ca2.91Na0.09Ti2-xRhxO7 (x?=?0.00, 0.02, 0.04, 0.06) were synthesized by a conventional solid-state reaction. Room temperature ferroelectricity has been confirmed. The remanent polarization increases with an increase of Rh content, which is due to a larger oxygen octahedral distortion by Rh doping. The coercive field increases with Rh doping as the pinning effect of oxygen vacancies reduce the mobility of domain wall. Remanent polarization and coercive field are caused by different mechanisms, so it is possible to modulate them independently to meet the requirement of application in ferroelectric field. The concentration of oxygen vacancy increased with Rh doping, leading to the significant increase of leakage current density. The bandgap of samples doped with Rh drastically decrease and the visible light response of the sample was improved by Rh doping due to the formation of impurity energy levels within the band gap.
The paramagnetic Meissner effect (PME) was observed on the field-cooled M-T curvesof melt-textured Y0.99Nd0.01Ba2Cu3O7-y in applied magnetic fields from 1kOe toJ0kOe. In the magnetic field below 1kOe, the sample undergoes the typical transi-tion from the normal state to the superconducting state when it was cooled down. Theapplied magnetic fields in which the PME occurs, are very high compared with thefields reported in literatures. We propose that the occurrence of the PME may be dueto a phase transition of flux vortices in the superconductors. 相似文献
An experimental-numerical method for measuring dynamic crack propagating velocities under stress wave loading is established in this paper. The experiments of the three-point bend specimen are done on the improved Hopkinson bar. Deflection of loading point, dynamic load and instantaneous crack length are measured, then crack propagating velocities are calculated. Experiments on 40Cr steel show that the results given by this method have a good agreement with that obtained by the resistance fracture gage method. Therefore this method is feasible for measuring crack propagating velocities under high loading rate and will have wide application. 相似文献