66 results about "Noise enhancement" patented technology

This invention incorporates the techniques of geophysical technology into medical imaging. Ultrasound waves are generated from multiple, simultaneous sources tuned for maximum penetration, resolution, and image quality. Digitally recorded reflections from throughout the body are combined into a file available for automated interpretation and wavelet attribute analyses. Unique points within the object are imaged from multiple positions for signal-to-noise enhancement and wavelet velocity determinations. This system describes gaining critical efficiencies by reducing equation variables to known quantities. Sources and receivers are locked in invariant, known positions. Statistically valid measurements of densities and wavelet velocities are combined with object models and initial parameter assumptions. This makes possible three-dimensional images for viewing manipulation, mathematical analyses, and detailed interpretation, even of the body in motion. The invention imposes a Cartesian coordinate system on the image of the object. This makes reference to any structure within the object repeatable and precise. Finally, the invention teaches how the recording and storing of the received signals from a whole body analysis makes a subsequent search for structures and details within the object possible without reexamining the object.

A method for displaying information relating to the range and Doppler of a remote target includes transmitting electromagnetic energy toward the target, and receiving reflected signals defining a two-dimensional (range-Doppler) radar image. The reflected signals are matched-filtered, which tends to blur the image. The image is deblurred while the features of thermal noise enhancement and irregularity of the deconvolved output are constrained to produce a single point deblurring output.

A diversity transmission scheme uses a number of antennas that is greater than the limitation of two transmitting antennas in the well-known Alamouti scheme. In an embodiment comprising four antennas, the antennas transmit in pairs such that each antenna transmits a block that is used in the Alamouti scheme. This increases the transmission rate. For example, the transmission of two signals at a given time slot increases transmission rate by a factor of two. The invention not only increases the number of antennas, but also increases the transmission rate. At the receiver end, the code is decoded without matrix inversion and without much noise enhancement. Moreover, noise enhancement stability is increased by a simple, partial interference cancellation scheme, that results in improved decoding performance.

An apparatus and method are disclosed for improving imaging based on a time-series of images. In one embodiment, a time-series of images are acquired using a same imaging protocol of the same subject area, but the images are spaced in time by one or more time intervals (e.g, 1, 2, 3 . . . seconds apart). A sub-region is projected across all of the images to perform a localized analysis (corresponding X-Y pixels or X-Y-Z voxels are analyzed across all images) that identifies temporal components within each sub-region. In some of the sub-regions, the temporal components are removed when the amplitude of the component is below a predetermined amplitude threshold. The images are then combined using the sub-regions with reduced components in order to obtain a single image with reduced noise.

A noise enhancement system for an automobile utilizes a single input of engine speed to derive a control signal used to drive a speaker. The speaker generates sounds that enhance or replicate the desired sounds from the engine. The output from the speaker is crafted dependent on the vehicle operating conditions. Instead of multiple transducers and inputs the noise enhancement system of this invention utilizes the single input of engine RPM engine speed. By utilizing only engine speed, the noise enhancement system is cost effective and significantly less complex.

The invention discloses an anti--sharpening masking and NSCT algorithm based mine image enhancing method. According to the invention, an anti--sharpening masking method is combined with NSCT (Non Subsample Contourlet Transform). The method includes dividing an image into three detail degrees including a high detail degree, a middle detail degree and a low detail degree. Low details are subjected to weighing replacement median filtering treatment. High detail areas are subjected to middle-degree enhancement and middle detail areas are subjected to comparatively high degree enhancement. An NSCT-based high frequency image enhancement algorithm is applied to the image and high frequency coefficients are classified based on a bayes threshold value estimation method. Strong edges, weak edges and noises are determined and different coefficients are treated separately. According to the invention, image fuzziness caused by noise removal is avoided and the enhancement of the image meets human eye vision features. Underground coal mine image features of low luminance and low contrast are improved and overshoot is avoided. Image detail loss is avoided, the enhancement effect is good and noise enhancement is inhibited.

The invention relates to a rotating-machinery life-stage identification method based on a deep self-encoding learning network of noise enhancement samples. For the purpose that extraction and expression of rotating-machinery life features as well as life stage identification are automatically learned under the condition of a small sample size, noise enhancement are conducted on training samples; after a plurality of sparse self codes are stacked, classification layers are added to construct the deep sparse self-encoding learning network which can not only automatically learn extraction of the life features, but also intelligently identify the life stages. Stepwise non-supervision adaptive learning and supervision fine tuning are conducted on the the samples obtained after noise enhancement through multi-layer sparse self encoding, so as to inhibit deep-network over fitting and improve network robustness. Therefore, automatic extraction and expression of the rotating-machinery life features are achieved, and finally intelligent identification of the rotating-machinery life stages in the classification layers are completed. The rotating-machinery life-stage identification method can be applied in identifying rolling bearing life stages, and identifying results are good under the condition of a small sample size.

An image of which restoration degree is determined based on the difference between each pixel and a neighboring pixel of the pixel in a light receiving element array 101 is divided into a plurality of areas based on the determined restoration degree, and the areas of the image are each restored in accordance with the determined restoration degree. By doing this, not only ringing and noise enhancement can effectively be suppressed but also the border between the areas which is conspicuous when the image is divided into two kinds of areas becomes less conspicuous, so that a natural restored image can be obtained.

The invention relates to a fabric defect detection method based on a deep learning algorithm, and the method comprises the following steps: carrying out the noise filtering and noise enhancement of a fabric original image through employing the image preprocessing technology; carrying out the decomposition and reconstruction of the fabric image after preprocessing through a Daubechies wavelet transformation method, carrying out the feature extraction of images in horizontal, vertical and diagonal directions after reconstruction, and forming a feature vector of the images; Carrying out the analysis and processing of the extracted feature vector of the fabric image through employing a multi-hidden-layer BP neural network model, and achieving the recognition and classification of defects in the fabric image. The method can automatically recognize the defects, and classifies the recognized defects.

A high dynamic range (HDR) compression method and apparatus modeled after the heat equation describing temperature changes in a thin plate. This approach allows combining high dynamic range compression together with noise reduction in a single process, to be performed within the same iteration of the heat equation. Noise reduction is of particular concern while performing HDR compression because brightening of dark areas during high dynamic range compression has the potential to increase noise levels. Performing image processing techniques in combination according to the invention provides enhanced results while lowering the overall processing overhead. This innovation extends the heat equation analogy by adding anisotropic diffusion as an additional term, which allows joint operation of HDR and NR and mitigates noise enhancement within HDR compression during shadow enhancement.

Spectral encoding methods are more robust when used with improved weak signal detection and synchronizations methods. Further robustness gains are achieved by using informed embedding, error correction and embedding protocols that enable signal to noise enhancements by folding and pre-filtering the received signal.

A system for receiving signals (e.g., optical signals) includes an input device, an amplification device, and a feedback device. The amplification device receives a signal from the input device and includes a transimpedance portion. The transimpedance portion includes a first section having a plurality of elements (e.g., resistors and transistors) and a second section having a plurality of elements (e.g., resistors and transistors). One or more of the elements (e.g. transistors or resistors) in the first and second sections are mismatched to introduce a systematic offset in the transimpedance stage, to make the net input referred offset of the amplification device unidirectional. The feedback device (e.g. an integrator) is coupled to an output of the amplification device and an input of the transimpedance portion to provide a unidirectional offset correction to the amplification device for reduced noise enhancement.

A method, apparatus, and computer usable program code for signal-to-noise enhancement. In one illustrative embodiment, a system is provided for signal-to-noise enhancement. The system includes a rank order filter and a comparator. The comparator receives an ordered set of data values from the rank order filter. A signal averager averages a subset of data values selected from the ordered set of data values by the comparator.

The invention relates to a color enhancement method suitable for remote sensing images, relates to the remote sensing image processing technology field and solves problems of limitation, poor robustness, poor image enhancement effect and noise enhancement existing in a remote sensing image color enhancement method in the prior art. The method comprises steps that to-be-processed remote sensing images and color images are inputted; color migration processing for the to-be-processed remote sensing images is carried out; deblocking processing of the images after color migration is carried out; fuzzy kernels of the images after deblocking processing are calculated; Wiener filtering is utilized for the deblocked images for deblurring; the images are subjected to Laplacian sharpening after deblurring, and the final result images are obtained. The method is advantaged in that color richness of the remote sensing images can be substantially improved, the image contrast is improved, the edge information is strengthened, and the enhanced images have the good visual effect.

The embodiment of the invention provides a training method of a child speech recognition model. The method comprises the steps of obtaining training data; training through a baseline acoustic model toobtain an unconditional generative adversarial network; inputting the random noise data into an unconditional generative adversarial network to obtain noise enhanced acoustic features; inputting thenoise enhancement acoustic features into a baseline acoustic model to obtain a posterior probability soft label corresponding to each frame of noise enhancement acoustic features; and training the children speech enhancement acoustic recognition model at least by taking the noise enhancement acoustic features, the soft label, the children speech training data and the hard label as sample trainingdata. The embodiment of the invention further provides a training system of the child speech recognition model. According to the embodiment of the invention, the pronunciation nature of the child speech is changed under the condition that the child speech is limited, diversified child speeches are generated, and the recognition accuracy of the child speech recognition model is improved.

The invention belongs to the technical field of infrared thermal imaging, A method for adaptive dynamic range transformation of infrared image is provided in detail. According to that histogram equalization process of the original data, a suitable starting and ending range of the data is determined, linear mapping and platform histogram equalization map are respectively carried out on the data within the range, and the gray level out of the range is truncated. It not only effectively solves the problem that the gray contrast is reduced and the image details are easily lost when the traditionallinear mapping method is used in image processing, At the same time, the problem of noise enhancement in processing infrared image with large temperature difference can also be guaranteed, which makes the scene imaging effect always keep strong contrast and detail, and suppresses the amplification of noise.

The invention relates to an identity authentication audio watermarking algorithm based on deep learning. The identity authentication audio watermarking algorithm is characterized by comprising the steps: 1) performing face segmentation, mute removal and spectrum conversion preprocessing on a data set; 2) training a designed identity watermark generation model to extract identity features of a speaker from an audio; 3) training a designed watermark embedding-extracting combined model to self-adaptively complete embedding and extracting of the watermark; 4) selecting proper weight ratio parameters through experiments, and adding robustness of a noise enhancement algorithm; and 5) visually completing identity authentication of the speaker. According to the invention, the dynamic identity authentication watermark information is generated from the speaker audio based on a generative adversarial model; the identity watermark is embedded and extracted based on an auto-encoder, and finally, identity authentication is carried out in a visual dynamic effect self-adaptive embedding and extracting mode different from a traditional static information and manual design scheme, so that the security of the audio information is ensured.

The invention discloses a noise enhancement-based linear minimum mean square error estimation method, and belongs to the field of signal processing. The method is a linear estimation method which combines noise enhancement with a linear minimum mean square error estimation method. The method comprises the following steps of: firstly adding independent additive noise for a nonlinear system input signal, and obtaining a noise added nonlinear system output signal through a nonlinear system; carrying out linear minimum mean square error estimation on an input parameter by utilizing the nonlinear system output signal, and establishing a noise enhanced parameter estimation model; and finally solving optimum additive noise under the model and obtaining parameter estimation under the optimum additive noise. According to the method, noise enhancement and a linear minimum mean square estimation method are combined, and through adding noise for the nonlinear system input, the aim of further decreasing the minimum mean square error generated when linear estimation is carried out on the input parameter by the system output signal is achieved.

The embodiment of the invention discloses a method for noise enhancement post-processing in speech decoding. The method comprises the following steps: the noise factor is obtained according to the spectral envelope and the long-time mean spectral envelope of a speech frame; whether the noise factor is higher than the preset threshold value or not is judged; the energy of the fixed code vector of the speech frame is smoothed according to the noise factor and the adaptive codebook gain of the speech frame; and the smoothed fixed code vector is obtained according to the energy of the smoothed fixed code vector. The embodiment of the invention also discloses a noise enhancement post-processing device. By adopting the method and the device, the complexity of calculation of the noise enhancement post-processing in the speech decoding can be lowered.

An equalizer includes n processing paths arranged to process blocks of data. N−1 data shifters are arranged so that each of the n−1 data shifters is in a corresponding one of the n processing paths and so that one of the n processing path has no data shifter. N finite filters are arranged so that each of the n finite filters is in a corresponding one of the n processing paths, and so that each of the n finite filters applies a corresponding set of finite filter coefficients to the blocks of data. Ghosts of the blocks of data are not eliminated as a result of the application of the sets of finite filter coefficients corresponding to the n finite filters, and n>2. An adder is arranged to add outputs from the n processing paths such that the addition eliminates ghosts of the blocks of data.

A frequency demodulator comprises a frequency discriminator configured to generate a frequency modulation signal from frequency modulated signal, circuitry for generating a phase modulation signal from the frequency modulation signal, and a click reduction signal processing (CRSP) circuit operable to remove noise enhancements from the phase modulation signal caused by clicks. By first converting the frequency modulation signal to a phase modulation signal, noise enhancements caused by clicks are more readily distinguished from other noise in the phase modulation signal. After the noise enhancements have been removed by the CRSP, the frequency modulation is recovered substantially free of clicks. Removal of the clicks results in an improved output signal-to-noise ratio, thereby advantageously extending the onset of the threshold effect.