71 results about "Log domain" patented technology

Servo control using ferromagnetic core material and electrical windings is based on monitoring of winding currents and voltages and inference of magnetic flux, a force indication; and magnetic gap, a position indication. Third order nonlinear servo control is split into nested control loops: a fast nonlinear first-order inner loop causing flux to track a target by varying a voltage output; and a slower almost linear second-order outer loop causing magnetic gap to track a target by controlling the flux target of the inner loop. The inner loop uses efficient switching regulation, preferably based on controlled feedback instabilities, to control voltage output. The outer loop achieves damping and accurate convergence using proportional, time-integral, and time-derivative gain terms. The time-integral feedback may be based on measured and target solenoid drive currents, adjusting the magnetic gap for force balance at the target current. Incorporation of permanent magnet material permits the target current to be zero, achieving levitation with low power, including for a monorail deriving propulsion from the levitation magnets. Linear magnetic approximations lead to the simplest controller, but nonlinear analog computation in the log domain yields a better controller with relatively few parts. When servo-controlled solenoids provide actuation of a pump piston and valves, electronic LC resonance measurements determine liquid volume and gas bubble volume.

The invention discloses a tone mapping method for a high dynamic range image. The method comprises the following steps of: firstly, reading the high dynamic range image, and acquiring a log-domain brightness image of the high dynamic range image; secondly, estimating illumination components of the high dynamic range image by multi-scale weighted least squares (WLS) filtering to avoid the phenomenon of halo generated at a high light and shade contrast place; thirdly, recovering brightness components of the high dynamic range image by using a multi-scale Retinex method, eliminating the gray of a 1% part of each of two poles to prevent under-exposure and over-exposure, and adjusting the brightness degree of the image according to a self-adaptive adjustment function; and finally, recovering the color of the high dynamic range image by introducing an index control saturation method. The tone mapping method for the high dynamic range image is applicable to display and output of the high dynamic range image on low dynamic range equipment, generation of the halo can be avoided, and the displayed image has the characteristics of high contrast and good detail reproduction effect.

The present invention relates to a vehicle audio identification method based on independent component analysis, and belongs to the audio processing and mode identification technology field. The utility model has the steps that audio frequency spectrums of different vehicles are gotten by a pretreatment of different vehicle sounds, frequency spectrums are trimmed down according to characteristics of the vehicle sounds, and coefficients of frequency domain are changed to log domain, thereby enhancing the robustness of the audio identification of the vehicle. Getting the characteristics of the vehicle sounds by using the independent component analysis adapts to small sample training characteristics of vehicle identification. The samples which are ready for identifying are rebuilt in a characteristic space consisting of independent components, a Euclidean distance between the samples which are ready for identifying and the center of the vehicle is gotten, and the samples which are ready for identifying are classified according to the Euclidean distance. The present invention can identify the vehicle sounds highly efficiently and fleetly, thereby especially adapting to real time computation circumstance. The present invention can be used in a plurality of circumstances like detecting of passing vehicles in military field and intelligent traffic system in civilian field and so on.

Decoding LDPC (Low Density Parity Check) code and graphs using multiplication (or addition in log-domain) on both sides of bipartite graph. A novel approach of decoding LDPC coded signals is presented whereby edge messages may be updated using only multiplication (or log domain addition). By appropriate modification of the various calculations that need to be performed when updating edge messages, the calculations may be reduced to only performing product of terms functions. When implementing such functionality in hardware within a communication device that is operable to decode LDPC coded signals, this reduction in processing complexity greatly eases the actual hardware's complexity as well. A significant savings in processing resources, memory, memory management concerns, and other performance driving parameters may be made.

The invention relates to a vehicle detecting algorithm in an intelligent traffic system, in particular to a vehicle detecting algorithm based on intrinsic image decomposition, which is universal at day and night. The specific practice thereof comprises the steps of: at first, carrying out derivation and filter to an initial input image on a log-domain to convert the initial input image to a gradient domain; then calculating the difference between the gradient map of a current image and the gradient map of a background image on the gradient domain to obtain the gradient map of a moving foreground image; using the method of intrinsic image decomposition to process the gradient map of the moving foreground image to obtain the gradient map of a photogram and the gradient map of a target image; extracting the pixels with the gradient amplitude larger than a threshold T in the gradient map of the target image and using the pixels as moving target points; and aiming at a certain region, considering that the vehicle passes through the region if the number of the moving target points exceeds a certain proportion of the total number of the pixels in the region. The vehicle detecting algorithm based on intrinsic image decomposition has the beneficial effects of removing the effects of shadows and illumination and being capable of accurately carrying out detection to the vehicles at day and night in real time.

Decoding LDPC (Low Density Parity Check) code and graphs using multiplication (or addition in log-domain) on both sides of bipartite graph. A means for decoding LDPC coded signals is presented whereby edge messages may be updated using only multiplication (or log domain addition). By appropriate modification of the various calculations that need to be performed when updating edge messages, the calculations may be reduced to only performing product of terms functions. When implementing such functionality in hardware within a communication device that is operable to decode LDPC coded signals, this reduction in processing complexity greatly eases the actual hardware's complexity as well. A significant savings in processing resources, memory, memory management concerns, and other performance driving parameters may be made.

A method of processing a layer in a convolutional neural network includes reading a plurality of different subsets of pairs of input vector values and corresponding weight vector values from an on-chip memory, storing the pair values in registers of a plurality of multiplication units, simultaneously processing the different subsets of pairs of values in the registers by multiplying the pairs of values and accumulating the results of the multiplications to obtain an output vector for a current layer. The input vector and the weight vector have floating point values and multiplying the pairs of values comprises adding corresponding pairs of integers in a log domain, and an input value or weight value is released from the on-chip memory after said input value or weight value is no longer needed for calculating non-computed output vector values

The present invention is directed towards systems and methods for transmitting a DNS beacon for storage. A method according to one embodiment comprises generating one or more data items for logging and transmitting a DNS request to an authoritative DNS server for a logging domain, with the host name of the DNS request comprising the one or more data items for logging. The DNS request, which comprises the data items for logging, is stored on a storage device.

The invention discloses a log domain analog wavelet transform circuit for detecting an electrocardiogram signal QRS wave, comprising a forming circuit of differential input current signals and a current mirror circuit, four parallel log domain differential voltage filter circuits, four log differential compression circuits and four exponential differential expanded circuits. Each input of the log domain differential voltage filter circuits is connected with the forming circuit of differential input current signals and the current mirror circuit through the log differential compression circuits. Each output of the log domain differential voltage filter circuits is connected with an input of the exponential differential expanded circuits. The log domain analog wavelet transform circuit has the advantages of being integrated in a piece of analog integrated circuit, real-time realizing continuous wavelet transform under the characteristic dimension of the electrocardiogram signals with low power consumption, satisfying application requirements of an implanted artificial cardiac pacemaker for low power consumption and small size.

Techniques for performing automatic gain control (AGC) at a wireless receiver are described. The total gain for the wireless receiver is achieved with discrete gain steps for analog circuitry and continuous gain for a digital variable gain amplifier (DVGA). An AGC loop is updated based on power measurements for an output signal from the DVGA. A first gain for the analog circuitry is selected from among multiple discrete gain values based on the AGC loop to maintain the average power of a baseband signal within a predetermined range at an analog-to-digital converter (ADC) input. A second gain for the DVGA is selected based on the AGC loop to maintain the average power of the output signal at a reference power level. The first gain is switched in a manner to avoid saturation of the ADC caused by the baseband signal and to provide switching hysteresis. The AGC may be performed in log domain and with multiple modes.

The invention discloses a method for unifying face illuminations on pictures with different size characteristics. Firstly, a log-domain total variation model is adopted to decompose an original face picture into small size characteristic pictures and big size characteristic pictures; then an illumination processing is carried out to the big size characteristic pictures which are greatly impacted by illumination changes and a minimal value filtering wave processing with threshold value is carried out to the small size characteristic pictures; finally, the processed pictures with different size characteristics are composed to obtain a face picture with unified illumination. The invention mainly carries out the illumination unification on the big size characteristic pictures which are greatly impacted by illumination changes to avoid the impact on face identification rate caused by changing the small size characteristics with no illumination change. Moreover, the invention does not give up the big size characteristics which are greatly impacted by illumination so as to avoid the identification information lack caused by just adopting the small size characteristics for face identification. The method of the invention can be realized easily without strict alignment to the face pictures and without any training samples, meets various practical application requirements.

The invention discloses a LBP (Local Binary Pattern) face recognition method for eliminating illumination evenness, and solves a problem that a LBP face recognition rate is affected due to the illumination evenness. The invention adopts a technical scheme that a face grey level picture I is changed to a Log domain to obtain a face picture R with illumination invariability information by separation; the face picture I is subjected to threshold value segmentation to obtain a picture M marked with an illumination area and a shadow area; according to the picture M, combining the picture R and the picture I into one new picture N without the illumination area and the shadow area; extracting the LBP textural feature vectors of each subblock of the picture N; and finally, connecting the LBP textural feature vectors of each subblock in series in sequence to obtain a feature vector, wherein the vector is the LBP feature vector which is finally extracted for eliminating illumination influence. The influence on the face recognition by illumination is greatly reduced, the illumination robustness of the face recognition is effectively improved, and face recognition accuracy is improved.

The invention relates to a training method for a language recognition model and a language recognition method. The language recognition method comprises the steps: extracting the phoneme posterior probability of speech data, converting the phoneme posterior probability into a log domain, conducting dimensionality reduction and mean and variance normalization, and then obtaining phoneme associated features; calculating Baum-Welch statistical magnitude by means of the phoneme associated features, and extracting a phoneme variance factor through the Baum-Welch statistical magnitude; modeling the phoneme variance factor, and establishing an SVM model (a language recognition model); and marking the SVM model by a phoneme variance factor of to-be-recognized speech data, conducting mean and variance normalization on the score, performing linear discriminant analysis and Gauss back end normalization on the normalized score to realize score correction, and finally obtaining a recognition result. Compared with a conventional language recognition method, the language recognition method of the invention has the advantages that the calculation complexity is reduced, the language recognition performance is obviously improved, and the method is highly practical.

Techniques are provided for the implementation of dynamically biased circuits. In these circuits, bias currents are varied according to signal amplitude. Benefits include reduced power dissipation, reduced noise, and increased dynamic range. The techniques can be employed in various types of circuits such as, for example, amplifiers, log-domain circuits, and filters.

The embodiment of the invention provides a signal detection method and device applied to a non-orthogonal multiple access system. The method and device are applied to a multi-user detector. The method comprises the following steps: according to an evaluation probability predicted value on a log domain, currently saved by another user node connected with a resource block node to be updated currently, determining an evaluation probability predicted value on the log domain of the resource block node to be updated currently, transferred to the user node, and sending the value to the user node so as to be saved; according to the evaluation probability predicted value on the log domain, currently saved by the user node and a prior probability value on the log domain of the user node, determining an evaluation probability predicted value on the log domain, transferred to a next resource block node connected with the user node, of the user node; and sending the value to the next resource block node connected with the user node so as to be saved. By applying the embodiment of the invention, the convergence rate of an MPA algorithm can be effectively increased, and the processing delay of the multi-user detector can be reduced.

The invention provides a voice signal processing method, device and system. The voice signal processing method comprises the steps of dividing a voice signal to be coded into subsignals according to the set quantity of sampled signals included by each of the subsignals; dividing the voice signal to be coded into voice frames according to voice energy of each subsignal in a log domain, wherein the voice energy of each subsignal included by each of the divided voice frames in the log domain is not less than a set value or less than the set value. Therefore, the voice signal with high communication meaning and the voice signal with low communication meaning can be distinguished well by the aid of high or low voice energy of each subsignal in the log domain. More coding bits are allocated for voice frames with high communication meaning (namely, a first type of the voice frames), fewer coding bits are allocated for voice frames with low communication meaning (namely, a second type of the voice frames), therefore, the quantity of the coding bits is decreased, compression efficiency in voice coding is increased remarkably, and transmission delay is reduced.

A polynomial kernel generator is configured to mitigate nonlinearity in a receiver path from a transmitter path comprising a nonlinear component in a communication device or system. The polynomial kernel generator operates to generate polynomial kernels that can be utilized to model the nonlinearity as a function of a piecewise polynomial approximation applied to a nonlinear function of the nonlinearity. The polynomial kernel generator generates kernels in a multiplier less architecture with polynomial computations in a log domain using a fixed number of adders.

Various embodiments of the present invention are related to methods and apparatuses for decoding data, and more particularly to methods and apparatuses for decoding data in a mixed domain FFT-based non-binary LDPC decoder. For example, in one embodiment an apparatus includes a message processing circuit operable to process variable node messages and check node messages in a log domain, and a check node calculation circuit in the low density parity check decoder operable to perform a Fast Fourier Transform-based check node calculation in a real domain. The message processing circuit and the check node calculation circuit perform iterative layer decoding.

The present invention is directed towards systems and methods for transmitting a DNS beacon for storage. A method according to one embodiment comprises generating one or more data items for logging and transmitting a DNS request to an authoritative DNS server for a logging domain, with the host name of the DNS request comprising the one or more data items for logging. The DNS request, which comprises the data items for logging, is stored on a storage device.

The invention discloses a remote-sensing image ship target segmentation method. The method carries out a filtering enhancement operation on an SAR image, and carries out contour extraction of a ship target on the basis of Randon transform. According to the method, texture details of the image are enhanced while image noises are reduced through using a filtering method of combining log-domain wavelet filtering and guided filtering; and a direction and a boundary of the ship target are also obtained, through using the Randon transform, to mitigate impacts of crosses and trailing on subsequent ship target identification.

The invention discloses an improved Retinex image defogging method. The method comprises the following steps of: inputting a haze image, estimating n(x,y) firstly and removing n(x,y) according to a new expression formula, wherein n(x,y) is estimated by use of gaussian filter because n(x,y) is caused by direct imaging of airglow, transforms slowly and belongs to a low frequency part; subtracting the estimated n(x,y) from an original drawing, normalizing the obtained values, and acquiring an intermediate result; carrying out gaussian filter on the intermediate result according to a single-scale Retinex algorithm, estimating an illumination component, and removing the estimated illumination component L(x,y) from the intermediate result in a Log domain; and carrying out automatic color gradation processing on the obtained result, eliminating an overexposure phenomenon in the result, and outputting the processed result. According to the improved Retinex image defogging method, the defogging effect is greatly improved, the distant scenery becomes clear, detail parts become obvious, the running speed is rapid and the real-time requirement can be satisfied.

The invention provides a voice signal reconstruction method and apparatus, wherein the method comprises: acquiring a sample voice signal and converting the same into a log-domain spectrum; extracting the content factor of each frame of sample voice signal by using a pre-trained content factor extraction model; extracting the speaker factor of each frame of sample voice signal by using a pre-trained speaker factor extraction model; training a content spectrum reconstruction model and a voiceprint spectrum reconstruction model based on the content factors and the speaker factors; inputting a target voice signal into the voiceprint spectrum reconstruction model to obtain a voiceprint spectrum in the target voice signal, and inputting the target voice signal into the content spectrum reconstruction model to obtain a content spectrum in the target voice signal; and adding the content spectrum and the voiceprint spectrum to obtain the reconstructed log-domain spectrum of the target voice signal. The voice signal reconstruction method and apparatus can decompose and reconstruct the voice signal based on the content and the speaker, has stronger correlation with tasks, and has high reconstruction efficiency. The reconstructed voice signal is relatively similar to the original voice signal.

The invention relates to a tune energy and human ear frequency selectivity based squeal judgment method for speed reducers / changers. The method comprises the steps: firstly, sampling noise signals ofa speed reducer / changer, carrying out A weighting on the noise signals, converting processed signals to log domains, and carrying out Fourier transformation, so as to obtain energy-amplitude frequencycurves; then, calculating feature loudness and critical band rate of the noise signals, calculating frequency band acuity of each feature loudness, and subjecting frequency band acuity to normalizingtreatment; subjecting the energy-amplitude frequency curves to frequency interval Gauss smoothing processing to obtain background noises, and selecting candidate tunes; calculating a frequency band acuity weighting factor according to a corresponding relationship between the critical band rate and frequency, and carrying out weighting, so as to obtain the energy-amplitude frequency curves of thenoise signals; calculating tune energy in bandwidth of the tunes, background noise energy and difference between the tune energy and the background noise energy, and carrying out primary and secondarysequencing on abnormal noise frequencies affecting squeal. According to the method, a low-frequency region becomes a non-sensitive region, and a high-frequency region highlights acuity attribute, andthus, a subjective sensation of human ears to squeal is effectively simulated.

The VEA Detector and Dynamic Range Controller of the invention more accurately measure constant or non-constant, periodic or aperiodic, signals and the use of such measurements to control the upstream and / or downstream processing of program signals, including without limitation audio, video, and power program signals. The invention uses an antilog module acting within the context of a log domain circuit such that the “averaging” at an integrator is linear, not logarithmic. However, since the detection is within the log domain, the dynamic range of the VEA Detector is exponentially larger.

A method and apparatus receiving number and using instruction to create resulting number approximating one of square root, reciprocal, or reciprocal square root of number. The resulting number as a product of that process. Using resulting number in a graphics accelerator. Apparatus preferably includes log-calculator, log-arithmetic-unit, and exponential-calculator. At least one of log-calculator and exponential-calculator include implementation non-linear calculator. The non-linear calculators may use at least one of mid-band-filter, outlier-removal-circuit. The invention includes making arithmetic circuit, log-calculator, log-arithmetic-unit and exponential-calculator. The arithmetic circuit, log-calculator, log-arithmetic-unit and exponential-calculator as products of manufacture. The arithmetic circuit may further include at least one of a floating-point-to-log-converter and / or a second of log-calculators. The arithmetic circuit may further include third and fourth log-arithmetic-units providing altered log domain numbers to third and fourth exponential calculators.

Techniques are provided for spatially adaptive tone mapping with dynamic brightness control. A methodology implementing the techniques according to an embodiment includes converting luminance data, from a received HDR image, to a logarithm domain and decomposing the converted data into a base layer, and one or more detail layers. The method also includes adjusting the layers, through shifting and scaling of the base layer and scaling of the detail layers, to map the amplitude data into a selected range. The method further includes converting the adjusted layers from the logarithm domain to a linear domain and calculating a tone compression parameter based on statistics of the adjusted log domain base layer. The statistics are associated with brightness of the received image. The method further includes applying a global tone compression function, based on the tone compression parameter, to the converted linear domain base layer to generate compressed linear domain layers.

The invention discloses a face characteristic extraction method, which comprises the following steps of 11, inputting a face photo; 12, performing log transformation on the face photo to obtain a log-domain image; 13, decomposing the log-domain image in an EMD (empirical mode decomposition) way to obtain an EMD image; 14, extracting a high-frequency component from the EMD image as the estimate of a reflecting component to obtain a lighting-invariable face expression and an estimated lighting component. According to the face characteristic extraction method, the reflecting component and the lighting component of the log-domain image in a log space form a linear superposition relationship by performing log space transformation on the face photo, and the log-domain image is decomposed in the EMD way to extract the lighting-invariable face expression from the log-domain image, so that the lighting component can be effectively distinguished from the lighting component, and a better effect can be achieved.

A method of processing a layer in a convolutional neural network includes reading a plurality of different subsets of pairs of input vector values and corresponding weight vector values from an on-chip memory, storing the pair values in registers of a plurality of multiplication units, simultaneously processing the different subsets of pairs of values in the registers by multiplying the pairs of values and accumulating the results of the multiplications to obtain an output vector for a current layer. The input vector and the weight vector have floating point values and multiplying the pairs of values comprises adding corresponding pairs of integers in a log domain, and an input value or weight value is released from the on-chip memory after said input value or weight value is no longer needed for calculating non-computed output vector values.

The invention discloses a logarithmic moment-based generalized Pareto distribution parameter estimation method. The invention mainly aims to solve the problems of poor estimation accuracy and low execution efficiency of an existing parameter estimation method. According to the technical schemes of the invention, the method includes the following steps that: 1, sea clutter data samples are obtained through a sea monitoring radar; 2, the obtained clutter data samples are normalized according to clutter power; 3, the check estimator of data in the log domain is calculated through using the normalized clutter data samples; and 4, a distribution parameter is calculated through using the check estimator. With the method of the invention adopted, the estimation accuracy of a traditional Pareto distribution parameter estimation method can be improved. The method has high computation speed and can meet a requirement for the real-time processing of the signals of a radar system and can be applied to target detection under sea clutter background.