92 results about "Power exponent" patented technology

The invention discloses a new fully homomorphic encrypted ciphertext division implementation method, which can implement encryption on real numbers (data) with any size and execute ciphertext homomorphic division operation on two encrypted ciphertexts. The method comprises the following steps: when a user uploads the data, implementing polynomial coding on the data, and then carrying out fully homomorphic encryption, uploading the encrypted ciphertexts and a power exponent of the dividend to a cloud server, ensuring that the cloud server can directly execute division operation to any two encrypted ciphertexts that are uploaded to the server, and then enabling the server to return the calculation result to the user, and ensuring that the user decrypts the result by using a private key of the encryption algorithm to obtain a plaintext calculation result.

The invention relates to a method for measuring load thresholds of geotechnical packing accumulated deformation states under cyclic loading. According to the method, brick walls and sand bags are adopted to construct a filling model, so as to form a testing system along with loading equipment and a displacement sensor, and a cyclic loading test is performed on the filling model, so as to obtain a changing curve of accumulated deformation with loading times; then the changing curve of segmented single loading average deformation with the loading times is obtained through calculation, and a negative power function as shown in the specification is adopted to fit, so as to obtain a relation curve of a power exponent Pi value with the change of load levels; a function relation between the Pi value and the load levels is obtained by virtue of cubic polynomial regression, and the corresponding load levels when P is equal to 2, 1 and 0 are obtained through back calculation, so as to be used as the load thresholds delta t1, delta t2 and delta t3 for distinguishing different evolution states of the accumulated deformation, wherein an experimental basis is respectively provided for design and maintenance of ballastless and ballast track bed structures of railways by the delta t1 and delta t2. The method has the advantages of convenience in model manufacturing, low cost, simplicity in operation and clear distinguishing criterion, and the obtained load thresholds are accurate and reliable.

The invention discloses a novel method for correcting an error of a wide-area ionized layer, wherein one of monitoring stations placed in the middle is taken as a central station according to the distribution of the monitoring stations for the ionized layer, and the central station and two edge stations radiating outwards form a triangle which is taken as a triangle sub-area. The covering area of the monitoring station is divided into areas consisting of a plurality of triangle sub-areas so as to establish a triangle sub-area ionized layer model covering the whole network service area. The vertical ionized layer delay at three vertexes (the monitoring stations) can be obtained through resolving according to the triangle sub-areas in which puncture points of the user ionized layer are located, the data (VTEC) of the ionized layer at the puncture points can be calculated through a distance power exponent weighting method, and then the data can be converted to the content (TEC) of a station-star oblique ionized layer through obliquity factors. The method solves the problem of accurate correction for the ionized layer in regions of China according to the characteristic that a grid model is difficult to establish for the topography distribution in China.

The invention relates to a large-core-diameter bend insensitive energy transmitting optical fiber. The large-core-diameter bend insensitive energy transmitting optical fiber comprises a core layer and cladding layers and is characterized in that the core layer is distributed according to a gradually-changing refractive index, the distribution power exponent alpha ranges from 1.8 to 2.1, the maximum value delta1%max of the relative refraction index difference of the core layer ranges from 1.0% to 2.5%, the radius R1 of the core layer ranges from 95 micrometers to 500 micrometers, the inner cladding layer and the outer cladding layer are sequentially arranged on the core layer from inside to outside, the relative refraction index difference delta2% of the inner cladding layer ranges from -0.6% to -1.3%, the specific value of the radius of the inner cladding layer to the radius of the core layer satisfies the equation that w1=R2 / R1 and ranges from 1.05 to 1.2, the outer cladding layer is a pure quartz glass layer, and the specific value of the radius of the outer cladding layer to the radius of the core layer satisfies the equation that w2=R3 / R1 and ranges from 1.1 to 1.4. According to the large-core-diameter bend insensitive energy transmitting optical fiber, the light receiving capacity and the light condensation capacity of the optical fiber can be remarkably improved, the optical fiber has good transmission efficiency and can bear high-energy laser transmission, the design of the profile with the gradually-change refractive index is adopted for the core layer, and therefore the optical fiber has the self-focusing effect and has more advantages in practical application such as laser processing.

The invention discloses a video fusion performance evaluating method based on high-order singular value decomposition. The video fusion performance evaluating method based on the high-order singular value decomposition mainly solves the problem that the fusion performance of video images containing noise cannot be evaluated in the prior art. The video fusion performance evaluating method based on the high-order singular value decomposition comprises the implementation steps that two registered reference videos and a registered fusion video are input respectively; a four-order tensor is formed by the input videos, the high-order singular value decomposition is carried out on the four-order tensor, and respective space geometrical characteristic background images and respective time motion characteristic images are obtained; the time motion characteristic images are divided into motion target areas and noise areas through a thresholding method; then, different evaluation indexes are designed respectively to evaluate all the areas; lastly, an overall performance evaluation index is structured through power exponent multiplication, and thus evaluation on the whole fusion performance of the video images is achieved. The video fusion performance evaluating method based on the high-order singular value decomposition can effectively, accurately and objectively evaluate the fusion performance of the videos in a noise environment, and can be used for monitoring fusion video image quality.

The invention discloses a deep neural network hardware accelerator based on power exponent quantization, relates to a processor structure for deep neural network convolution calculation hardware acceleration, and belongs to the technical field of calculation, calculation and counting. The hardware accelerator comprising: an AXI-4 bus interface, an input cache region, an output cache region, a weight cache region, a weight index cache region, an encoding module, a configurable state controller module and a PE array. The input cache region and the output cache region are designed into a row cache structure; the encoder encodes the weights according to an ordered quantization set that stores absolute values of all quantized weights. when the accelerator performs calculation, the PE unit readsdata from the input cache region and the weight index cache region to perform shift calculation, and sends a calculation result to the output cache region. According to the method, floating point multiplication is replaced by shift operation, so that the requirements on computing resources, storage resources and communication bandwidth are reduced, and the computing efficiency of the acceleratoris further improved.

The invention discloses a human face recognition method and system based on kernel non-negative matrix factorization, and a storage medium. The method comprises the steps: constructing a fractional order inner product kernel function, wherein the fractional order inner product kernel function exerts no limit on a power exponent parameter; obtaining a fractional order inner product kernel non-negative matrix factorization algorithm through the combination of the fractional order inner product kernel function and the kernel non-negative matrix factorization; and carrying out the human face recognition through the fractional order inner product kernel non-negative matrix factorization algorithm. The method solves a problem that the power parameter of a polynomial kernel function just can be an integer, and enables the selection of the power parameter to be more flexible. The method effectively irons out the changes of posture and illumination in human face recognition. Moreover, the method is quick in convergence and excellent in recognition performance.

The invention discloses a remote sensing method for estimating a particle fraction structure of offshore water body phytoplankton. The method comprises the steps of performing diagnostic pigment analysis (DPA) on HPLC data which are obtained through voyage, and obtaining a Chla concentration value which corresponds with each particle diameter; performing derivation based on a power exponent relation according to field measured remote sensing reflectivity data for obtaining a localized parameter; and finally applying an established model to a satellite, thereby obtaining PSC time-space distribution information which is inversed by the satellite. The remote sensing method has advantages of settling a problem of offshore water body phytoplankton particle fraction structure inversion, algorithm shortage in China, expanding application of water color remote sensing satellite data, and improving acquiring capability for an essential parameter of phytoplankton particle fraction structure.

The invention relates to a method for measuring a load threshold value of a soil deformation state of a railroad bed under a static load. The method is characterized by comprising the following operating steps: building a filling model in a model case formed by rigid side walls and applying static loads in grade to the filling model by using a round rigid loading plate according to different proportionality coefficient lambda of the ultimate bearing capacity; measuring and calculating plastic deformation rates vi(t) of soil at different moments t in a loading process by using a displacement sensor, taking data points of the rates and fitting the data points according to a negative power function v(t)=At-alpha; and obtaining a power exponent alphai value corresponding to the load proportionality coefficient lambda, and fitting the power exponent alphai value to obtain an alpha-lambda curve according to a cubic polynomial, wherein taking loads piI=lambdaIsigmaf and piII=lambdaIIsigmaf corresponding to lambdaI and lambdaII of two curature maximum values of the alpha-lambda curve are adopted as the load threshold value of the soil deformation state. piI and piII can be used for providing test grounds for design and optimization of ballastless / ballast track beds of a high speed railway and a deformation state evaluation and regulation principal, respectively. The method has the characteristics of being relatively short in testing time and specific in criterion, and the obtained load threshold value is relatively accurate and reliable.

A decoder for performing log-sum corrections by means of a linear approximation, putting stress on speed, with a reduced circuit dimension without adversely affecting the decoding performance of the circuit. The decoder includes a linear approximation circuit that computes the log-sum corrections using the function F=−a P−Q+b, where the coefficient −a represents the gradient of the function and the coefficient b represents, the intercept and are expressed by a power exponent of 2.

A decoding method includes that when encoding at a sending terminal, for a m-order primitive polynomial P(x), a primitive field element in galois field GF(2m) is represented by α; a lookup table f(αj) for different power exponents of α is established, where the value of j is selected from all the integers ranging from 0 to 2m−1, with a total number of 2m; a generator polynomial G(x) is expanded to obtain a polynomial with respect to x, with coefficients being an addition or subtraction of the power exponents of α; a remainder polynomial R(x), obtained by dividing code word polynomial Q(x) by the generator polynomial G(x), is a polynomial with respect to x, with coefficients being an addition or subtraction of the power exponents of α; and the coefficients of the generator polynomial G(x) and the remainder polynomial R(x) are both calculated using data found in the lookup table f(αj).