35results about How to "Reduce multiplication" patented technology

The invention belongs to a video and image processing technology, in particular relates to a method for convolving a symmetrical convolution kernel of a video / audio signal by applying an SSE (Streaming SIMD Extension) instruction set. The method comprises the following steps of: dividing a floating-point signal sequence S(n) of an audio input signal into a left interval, a middle interval and a right interval and carrying out convolution processing on the left interval and the right interval by using a common C language algorithm; and carrying out convolution processing on a signal in the middle interval by utilizing the SSE instruction set according to one group of every four adjacent points. In the method, the symmetrical characteristic of the convolution kernel is utilized, and half multiplication calculation is effectively reduced; in addition, because the SSE instruction set can calculate four points once, the number of the instructions is effectively reduced, no any conditional transfer instruction exists in concrete processing, and the production line of a CPU (Central Processing Unit) can be more fully utilized.

The invention relates to a digital pulse amplitude analyzer based on symmetrical zero area trapezoidal shaping. The digital pulse amplitude analyzer comprises an ADC, a deconvolution machine, a trapezoidal synthesizer and a symmetrical zero area trapezoid synthesizer which are connected in sequence. According to the digital pulse amplitude analyzer, a symmetrical zero area trapezoid is synthesized based on a widely-applied trapezoidal shaping algorithm, through the shaping algorithm, a base line can restore stability, the symmetrical zero area trapezoid adopts a differential recurrence equation form algorithm, so that a large amount of multiplications needed in the convolution operation are reduced, an operation result can be output in real time, and the dead time of measurement of a system is eliminated.

The invention discloses a simplified weighted repeat pseudo-median filtering method with negative coefficients. The method comprises the following steps: taking the frequency response characteristics of an approximation same-order finite impulse response (FIR) filter as an objective, designing weighting coefficients containing negative numbers for weighted pseudo-median computation of filtering window data by an MALLOWS comprehensive method, adjusting the window data by the weighting coefficients; simultaneously, computing weighting coefficients based on the slopes of a group of special structures designed for the filter, computing the overall slope of the window data repeatedly by the weighted pseudo median, carrying out linear transformation on the window data by the slope; and finally carrying out weighted computation on the transformed data by the pseudo-median weighting coefficients to obtain the output value of weighted repeat pseudo-median filtering. The method realizes a better frequency selection characteristic than that of the existing negative coefficient weighted median or pseudo-median filtering method, and simplifies computation in the algorithm to meet the requirement of systems with finite resources, such as a wireless sensor and the like for low power consumption.

A neural network convolution operation device and method, wherein the device is used to realize the convolution operation of the weight matrix and neurons in the neural network in the form of matrix multiplication, which includes: a shift operator, used for the neural network The element matrix and the weight matrix are respectively subjected to winograd transformation to obtain a transformed neuron matrix and a transformed weight matrix; a matrix multiplication operator is used to multiply the transformed neuron matrix and the transformed weight matrix matrix multiplication operation to obtain a multiplication matrix; the shift operator is also used to perform winograd inverse transformation on the multiplication matrix to obtain a convolution operation result; the controller is used to control the shift operator to perform winograd transformation or The winograd inverse transform is also used to control the matrix multiplication operator to perform matrix multiplication.

The invention provides a soft demodulation method of 8PSK Gray mapping. The realization method is as follows: a hard decision line corresponding to each bit is found in a 2D coordinate plane, the minimum distance from a receiving signal to the hard decision line is taken as the size of the bit soft information, and a symbol of the soft information is determined by the position of the receiving signal in relation to the hard decision line.

The invention discloses a multi-view image fusion method based on a floating-point type cylindrical lens grating. Based on the prior art, for the coefficient array generation, firstly, multiplication and division are adopted and then complementation is carried out, and then the calculation of the rows and columns is deformed, therefore, the number of the complementation can not only be reduced, but also the multiplication of floating-point type data in the calculation after the deformation can be reduced, and the amount of calculation is less. On a formula after rows and columns deformation, the complementation is converted, there is no complementation for the floating-point type data in hardware, and a comparator converts the operation into an addition and subtraction capable of being realized by the hardware; and at the same time, when collecting image sub-pixels, each sub-pixel in a fusion image is derived from two different viewpoints, and each viewpoint accounts for a different proportion, thus sub-pixel acquisition and weighting are performed on multiple views and then the views are fused into a glass-free 3D image suitable for the display of the floating-point type cylindrical lens grating, thereby saving more hardware resources while guaranteeing the precision.