1026 results about "Smoothing" patented technology

A system and method for synchronizing a local clock to a reference clock using a linear model of the error between the local clock and the reference clock is disclosed. In one embodiment, a direct smoothing process is used in conjunction with the linear model to estimate a frequency offset by which the frequency of an oscillator of the local clock is adjusted. Also disclosed herein is a phased-lock loop (PLL) adapted to synchronize a local clock with a reference clock using the direct smoothing process, as well as a system implementing the PLL for timing the playout of data received from a transmitter.

A method of analyzing electrocardiogram (EKG) data for use in the diagnosis of heart disease, prognosis of cardiac conditions, and the monitoring of heart disease therapies is disclosed. The method utilizes a wavelet-based multifractal analysis with one or more of (1) a discrete wavelet smoothing step to remove the effects of abnormal beats; (2) “Levy flight” analysis to detect the frequency of abnormal beats known to adversely affect the multifractal (MF) analysis; and (3) MF alpha analysis, a multifractal extension of monofractal short term (ST) alpha analysis. The invention further comprises an EKG test battery comprising Levy flight anomalous beat/beat cluster screening, followed by (ST) MF alpha analysis and MF Holder analysis (when validated by the Levy flight analysis). The wavelet smoothing step can also be used to classify human EKGs by observing the effect of sequential smoothing on the MF Holder coefficient. Alternative choices to the wavelet smoothing approach to removal of abnormal beat effects include probability distribution function analysis to determine the MF Holder coefficient directly, abnormal beat ridge skeleton removal to remove the offending beats based on a direct multifractal spectrum calculation, and the calculation of various types of entropy coefficients for the EKG time series.

This is directed to animating transitions in a user interface as the orientation of the user interface changes. An electronic device can display a user interface in any suitable orientation, including for example in portrait and landscape modes (e.g., based on the orientation of the display). To provide a resource efficient and aesthetically pleasing transition, the electronic device can pre-render the interface in the final orientation, and define an animation by cross-fading the interface between the initial and final interfaces. In some embodiments, the electronic device can identify distinct regions of the interface, and define separate animations for each region. The separate animations can be overlaid and displayed simultaneously to provide a uniform animated transition between the initial and final interfaces.

An impulse radio communications system using one or more subcarriers to communicate information from an impulse radio transmitter to an impulse radio receiver. The impulse radio communication system is an ultrawide-band time domain system. The use of subcarriers provides impulse radio transmissions added channelization, smoothing and fidelity. Subcarriers of different frequencies or waveforms can be used to add channelization of impulse radio signals. Thus, an impulse radio link can communicate many independent channels simultaneously by employing different subcarriers for each channel. The impulse radio uses modulated subcarrier(s) for time positioning a periodic timing signal or a coded timing signal. Alternatively, the coded timing signal can be summed or mixed with the modulated subcarrier(s) and the resultant signal is used to time modulate the periodic timing signal. Direct digital modulation of data is another form of subcarrier modulation for impulse radio signals. Direct digital modulation can be used alone to time modulate the periodic timing signal or the direct digitally modulated the periodic timing signal can be further modulated with one or more modulated subcarrier signals. Linearization of a time modulator permits the impulse radio transmitter and receiver to generate time delays having the necessary accuracy for impulse radio communications.

The present invention provides an image processing method. In one embodiment, the method provides discontinuity-preserving image smoothing and segmentation, noise reduction, reduction of image variations particularly those variations caused by illumination conditions, exposure compensation, and dynamic range compression. The present invention also provides an electronic image sensor that is able to detect high dynamic range optical images and produce electronic images with reduced dynamic range.

A smoothing algorithm for round trip time (RTT) measurements is provided to a network device to effectively deal with variations or other potential anomalies that may occur in RTT measurements. The algorithm involves: first determining what should be considered a very high or a very small value for a RTT sample. If a new RTT sample is in an acceptable range, then the network device performs a relatively basic smoothing. If the new RTT sample is much higher than a current RTT value, then the network device ignores the value of this RTT sample a few times. If the network device still detects this large value after ignoring that value for some time, then the network device factors this value into the current RTT value using an additive increase. Similarly, if the value of the new RTT sample is much lower than current RTT value, the network device ignores the value of the new RTT sample a few times. If the network device still sees this small/low value after ignoring that value for sometime, then the network device factors this value into the current RTT value using a multiplicative decrease. An effective RTT value results, which can be used singly or in combination with other metrics to load balance network traffic.

The present invention relates to spatial prediction of pixels of a block, the block being a block of a digital image. In particular, for a block pixel at least one reference pixel(s) is selected and out of the selected reference pixel(s), the block pixel is predicted. In particular, the prediction of the block pixel is performed depending on the distance of this pixel to the reference pixel(s) from which it is to be predicted.

The invention discloses a video-based 3D human face expression cartoon driving method, which comprises the following steps: (1) image preprocessing, namely improving the image quality through light compensation, Gaussian smoothing and morphology operation of a gray level image; (2) BP human face detection, namely detecting a human face in a video through a BP neural network algorithm, and returning the size of the position of the human face for providing a smaller search range for human face characteristic point positioning of the next step to ensure instantaneity; (3) ASM human face characteristic point positioning and tracking, namely precisely extracting characteristic point information of human face shape, eyes, eyebrows, mouth and nose through an enhanced active shape model algorithm and a predicted meanshift algorithm, and returning the definite position; and (4) data information conversion, namely converting the data information acquired in the human face characteristic point positioning and tracking step to acquire the motion information of the human face. The method can overcome the defects in the prior art, and can achieve live human face cartoon driving effect.

The invention discloses an indoor AGV (Automated Guided Vehicles) path planning method based on an improved A* algorithm. The method comprises the following steps: firstly, constructing an environmental map by adopting a grid method, and expanding the grid map according to the size of an AGV; then, optimizing a search strategy of a traditional A* algorithm, introducing a multi-value map model, cost function evaluation and path inflexion smoothing, and adding safety factors as path evaluation indexes; and finally, adding a starting point and a target point of the AGV to the grid map, adopting an improved distance evaluation function, taking the node with the smallest cost function as a node of the next step, gradually planning a running route of the AGV, and finally displaying the running route in an electronic map. Simulated analysis verifies that the AGV path planning based on the improved A* algorithm is higher in search route safety factor, smoother in path and higher in search speed.

To achieve an improved convergence behaviour for, e.g., echo cancellation there is provided a digital adaptive filter which includes a filter coefficient update means to successively update filter coefficients in accordance with an input signal, an estimated power of the input signal, and an error signal between the input signal filtered in the digital adaptive filter and the input signal propagated along an external path being modelled by the digital adaptive filter. Here, an input signal power estimation means is adapted to perform recursive smoothing for an increasing input power and a decreasing input power in an asymmetric fashion with different smoothing factors. In case the estimation is carried out in the frequency domain the step size for the update of filter coefficients may be calculated for each frequency band individually.

Methods and apparatus are provided for visual background subtraction using one or more preprocessing modules. One or more effects are detected in a received image signal and one or more blocks are selectively enabled to preprocess the image signal to compensate for the detected one or more effects. Visual analysis is then performed on the preprocessed signal using background subtraction. A spatially-variant temporal smoothing of the image signal is also disclosed. The spatially-variant temporal smoothing can be achieved by the mixing of a new intensity value with a previous intensity time-average as determined by a weighting matrix. The mixing can be influenced by a dynamic bias term that is a real-time estimate of a variance at the pixel, such as a degree of change, and the weighting can be determined by a relative stability of an observed value compared to a stability of the time-average.

A video processing apparatus and methodology are implemented as a combination of a processor and a video decoding hardware block to decode video data by performing piecewise processing of overlap smoothing and in-loop deblocking in a macroblock-based fashion. With this approach, a smaller on-board memory may be used for the in-loop filtering operations of the video decoding hardware block. By pipelining the piecewise processing operations, latency in the filtering operations is hidden and the filtering output is smoothed, thereby avoiding the need for bursts of fetching and storing of blocks.

The invention discloses a human face three-dimensional reconstruction and human face replacing video editing system and method based on deep learning, and the system consists of a deep learning module, a three-dimensional reconstruction module, a video preprocessing module, a two-dimensional image generation module, and a video editing module. The deep learning module obtains necessary parameters for the generation of a three-dimensional model through a training convolution neural network. The three-dimensional reconstruction module obtains a corresponding three-dimensional model according to an inputted human face image/video. The video preprocessing module carries out the processing of a target video, and obtains the human face feature points and expression parameters corresponding to each frame. The two-dimensional image generation module enables a three-dimensional model of a target human face to generate a two-dimensional human face image corresponding to each frame. The video editing module enables a new human face image to replace the target image, and carries out the smoothing and adjustment of the illumination conditions.

The invention discloses an egocentric vision in-the-air hand-writing and in-the-air interaction method based on a cascade convolution nerve network. The method comprises steps of S1: obtaining training data; S2: designing a depth convolution nerve network used for hand detection; S3: designing a depth convolution nerve network used for gesture classification and finger tip detection; S4: cascading a first-level network and a second-level network, cutting a region of interest out of a foreground bounding rectangle output by the first-level network so as to obtain a foreground region including a hand, and then using the foreground region as the input of the second-level convolution network for finger tip detection and gesture identification; S5: judging the gesture identification, if it is a single-finger gesture, outputting the finger tip thereof and then carrying out timing sequence smoothing and interpolation between points; and S6: using continuous multi-frame finger tip sampling coordinates to carry out character identification. The invention provides an integral in-the-air hand-writing and in-the-air interaction algorithm, so accurate and robust finger tip detection and gesture classification are achieved, thereby achieving the egocentric vision in-the-air hand-writing and in-the-air interaction.

The invention discloses a speech enhancement method using a speech existence probability, after the processing of the invention, the speech quality is higher, and the noise reduction amount is larger.The method is realized by the following technical scheme that: on the basis of a MIS measure speech enhancement method, the speech existence probability is used, an inputted noisy speech is sampled,framed and windowed, and then a noisy speech spectrum is obtained through fast Fourier transform (FFT); the noise estimation of the obtained speech spectrum is carried out, based on a non-stationary noise minimum search algorithm of statistical information, the smoothing between front and back frame noise estimation values is carried out by using inter-frame correlation, and a noise power spectrumis estimated; at the same time, speech prior signal-to-noise ratio estimated values obtained in several front and back frames are smoothened; then the speech existence probability and a MIS measure gain factor are combined, the obtained noisy speech spectrum is multiplied by a new gain factor, and a spectrum of the enhanced speech is obtained; and the inverse fast Fourier transform (IFFT) is performed to obtain an enhanced time domain speech signal.

A method of making an organic light emitting device (OLED) is disclosed wherein an inert insulating conformal smoothing layer is deposited over a protruding structure. The smoothing layer is patterned to expose portions of the structure underlying the smoothing layer and defining active regions of the device. The inert smoothing layer is treated, preferably by heat reflow, to taper the layer over the stepped edges of the structure on the exposed portions. Additional layers are then deposited over the smoothing layer and the exposed portions of the structure. The smoothing layer blunts all underlying layer edges and provides sloped edges wherever a step occurs from one layer to another. This effect results in a homogeneous field across the pixel and the continuity in the layers deposited after the photoresist layer.

The invention discloses a tri-pole direct current transmission system topology structure based on an MMC. The tri-pole direct current transmission system topology structure is characterized in that direct current transmission is performed in a tri-pole connecting mode, and each pole comprises a rectification-side converter transformer, a rectification-side converter, a rectification-side smoothing reactor, a direct current transmission line, an inversion-side smoothing reactor, an inversion-side converter and an inversion-side converter transformer group. The rectification-side converters and the inversion-side converters are in modular multi-level structures, sub-modules of different amounts are connected in series and in parallel to be adapted to requirements for different voltage classes and currents, and three transmission lines used in a tri-pole direct current system are cables or overhead lines. The tri-pole direct current transmission system topology structure is high in reliability and capable of meeting the requirement for changing alternating current lines into direct current lines, phase change failures on the inversion side can be avoided under the condition that an alternating current system breaks down, voltages on two sides of the alternating current system can be stabilized, a alternating current filtering and reactive compensation device is not needed, and good application values are provided.

The invention relates to a mobile robot path planning method based on improved A-star algorithm, and belongs to the field of mobile robot path planning. The method is as follows: firstly performing global path pre-planning in a static obstacle environment; then performing secondary smoothing on path points of the global path pre-planning to obtain a pre-planned path; and finally adopting an artificial potential field method and a method for global path pre-planning in the static obstacle environment to perform local path planning when encountering a dynamic obstacle. The method is no longer limited to points in a grid. In a path-finding process, a current node and an extended node are connected, and the obstacle is detected by judging whether a route point connection line and an obstacle edge have intersections.

A combined initial alignment system and alignment method for a strapdown inertial navigation system of an underground coal mining machine belong to initial alignment system and alignment method for positioning navigation of underground coal mining machines. The system is composed of a strapdown inertial navigation system installed on the coal mining machine, a wireless sensor network mobile node, an inclination sensor, a geomagnetic field sensor and anchor nodes installed on a hydraulic support. After coarse alignment of the strapdown inertial navigation system, the wireless sensor network is used to detect the location information of the underground coal mining machine, the inclination sensor measures the roll and pitch angle, and the geomagnetic field sensor measures the yaw angle. A pose measurement equation of the coal mining machine is constructed, and the pose measurement equation is combined with an error model after the coarse alignment of the strapdown inertial navigation system to establish a state equation; fusion smoothing is carried out to obtain the accurate location information of the coal mining machine; and the precise alignment of the strapdown inertial navigation system is conducted to complete the initial alignment. The invention realizes precise initial alignment of strapdown inertial navigation combination under the severe closed condition in coal mine, and greatly improves the precision of the integrated positioning under large misalignment angle of the strapdown inertial navigation system of the coal mining machine.

Input MRI brain images are positioned so as to correct a spatial deviation, gray matter tissues are extracted from these images to effect a first image smoothing, the thus-obtained images are subjected to anatomical standardization, a second image smoothing is effected, the gray level is corrected, brain images after correction are statistically compared with MRI brain images of normal cases, thereby providing the diagnosis result. In this instance, the brain images are automatically checked for input images regarding the resolution dot density and the like, the result of gray matter tissue extraction and the result of anatomical standardization, by which specifications of input images and the like can be confirmed objectively and automatically to make a diagnosis automatically by image processing. Further, an ROI-based analysis is made to provide the analysis result as the diagnosis result.

Provided is an image processing apparatus that can reduce degradation in image quality due to heat haze and the like. An input image memory 110 stores an input image group 200 and an input image 210. A time-series smoothed image memory 112 stores a time-series smoothed image 220 obtained by time series smoothing. A moving body removed image memory 113 stores a moving body removed image 230 that is a background image. A histogram analyzing unit 130 computes a differential of histograms of partial images for the input image 210 and the time-series smoothed image 220 or the moving body removed image 230, and calculates similarity information between the images. On the basis of the similarity information, a region dividing unit 140 divides the input image 210 into a moving body region, and a region other than the moving body region. An image correcting unit 150 performs image synthesis of the input image 210 or the time series-smoothed image 220, for each of the regions.