35results about How to "Improve denoising effect" patented technology

This application discloses an image processing method and apparatus in the computer vision field of the artificial intelligence field. The image processing method includes: obtaining a channel image corresponding to a to-be-processed raw (raw) image, where the channel image includes at least two first channel images; performing image restoration processing on the first channel image to obtain the restored first channel image; and performing image restoration processing on the to-be-processed raw image based on the restored first channel image to obtain a raw image obtained through joint demosaicing and denoising processing. The technical solution of this application can improve image quality of the raw image obtained through joint demosaicing and denoising processing.

A method for enhancing steel surface images based on a regional autonomous strategy belongs to the field of computer vision and image processing technology. The steps are as follows: 1. Image preprocessing: Acquire the original image of the steel surface and convert the input RGB color image to a grayscale image; 2. Defect region detection: Use the Canny edge detection algorithm to perform edge detection on the grayscale image and extract potential defect boundary information; 3. Regional CLAHE enhancement: Based on the region mask generated in step 2, perform contrast-limited adaptive histogram equalization on the defect region and the background region respectively, and set different enhancement parameters; 4. Median filtering; 5. Weighted image fusion: Perform weighted fusion processing on the enhanced image obtained in step 3 and the filtered image obtained in step 4 to obtain the final output image. This invention can improve the overall quality and defect identifiability of steel surface images through the above method.

This invention discloses a partial discharge (PD) signal denoising method based on image information entropy and multivariate variational mode decomposition (MMD). The method involves converting the noisy signal into a grayscale image, calculating the image information entropy, and optimizing the number of modes K in the MMD by combining Pearson correlation coefficient and execution efficiency to determine the optimal value. The noisy PD signal is then decomposed. The kurtosis value of each intrinsic mode component is calculated, and the nature of the mode component is determined based on a threshold, classifying it as either a dominant PD component or a noise component. A mathematical statistical method using the 3σ criterion is employed to filter out normally distributed white noise. The reconstructed signal is then denoised using an improved wavelet thresholding method to obtain the denoised PD signal. This denoising method accurately reduces noise in noisy PD signals, achieving good noise suppression and restoring the waveform characteristics of the PD signal while maintaining high execution efficiency.

The invention discloses a blood oxygen signal denoising method and system based on adaptive filtering, and the method comprises the steps: synchronously collecting an original blood oxygen signal and a three-axis acceleration signal through a sensor module, and obtaining an original blood oxygen signal sequence and a three-axis acceleration signal sequence which are aligned in time; the three-axis acceleration signals are fused into one-dimensional motion artifact reference noise signals used for representing the limb motion state of the user; the method comprises the following steps: dynamically calculating an adaptive step length factor sequence through a nonlinear mapping function based on multilevel energy analysis of a motion artifact reference noise signal; constructing a time-varying step size adaptive filter, and calculating noise components in the original blood oxygen signal by using the motion artifact reference noise signal and an adaptive step size factor; and subtracting the calculation noise signal from the original blood oxygen signal, and reconstructing to obtain a pure blood oxygen signal without motion artifacts. According to the method, accurate calculation and efficient suppression of motion artifacts can be achieved, and the purification quality and processing efficiency of blood oxygen signals are remarkably improved.

The invention discloses a high-dimensional seismic data denoising method and device. The method comprises the following steps: segmenting a high-dimensional seismic data volume into a plurality of data blocks by using a sliding window; taking the data block with the corresponding de-noised data as a sample to obtain a sample set, and training a Transform model to obtain a de-noising model; and inputting a plurality of data blocks obtained by segmenting the high-dimensional seismic data volume into a denoising model, and obtaining a denoised seismic data volume of the high-dimensional seismic data volume according to an output result of the model. According to the method, intelligent suppression of high-dimensional seismic data noise can be realized by using high-dimensional features of seismic data based on a deep neural network Transform model, and the denoising precision is high.

The invention provides an image classification model training method and an image classification method. The method comprises the following steps: acquiring an image classification model; wherein the image classification model comprises a frozen visual coding network, a field prompt parameter corresponding to at least one first image generation field and a classification network; in response to monitoring of a newly added second image generation domain, freezing the visual coding network, and adding a learnable domain prompt parameter corresponding to the second image generation domain and a classification network in the image classification model; based on the first sample image and the second sample image, performing incremental training on a domain prompt parameter and a classification network corresponding to a second image generation domain; therefore, full retraining of the backbone network is avoided, generalization and adaptability of the model to the heterogeneous generation algorithm are improved, and efficient and stable detection performance can also be achieved under the scene that the newly generated model only has a small number of samples.

The application discloses a double-parameter HP filtering denoising algorithm and process for infrared spectrum signal processing, wherein upper and lower limits Lamda_1 and Lamda_2 of parameter scale estimation are given in advance, two filtered signals are obtained by performing traditional HP filtering respectively, the two filtered signals are introduced to comprehensively construct a new HP filtering expression, HP filtering is performed under the new comprehensive reconstructed HP filtering expression, and the parameter value when the expression takes a minimum value is taken as an output result in the range of [Lamda_1, Lamda_2] for optimization searching. The local extreme value is searched in the two parameter estimation ranges, and HP filtering is performed under the new expression. Since the application constructs the HP filtering expression of the comprehensive parameter upper and lower limits and performs the optimization processing in the parameter upper and lower limits, the scale estimation of the signal noise can be better, the problem that the uncertain selection of the parameter in the HP filtering denoising process leads to the difficulty in grasping the scale of the high-frequency noise denoising can be solved, and better denoising effect can be obtained.

The invention relates to the technical field of illegal cooking oil identification, in particular to an illegal cooking oil identification method, device and equipment adopting illegal cooking oil CatBoost. The method comprises the following steps: acquiring Fourier infrared spectrograms of different types of illegal cooking oil; preprocessing the collected fourier infrared spectrum of the illegal cooking oil through self-adaptive wavelet transform, and constructing a fourier infrared spectrum set of the illegal cooking oil; extracting swill-cooked dirty oil feature information in the swill-cooked dirty oil Fourier infrared spectrum set by adopting a multi-layer RNN model, and constructing a swill-cooked dirty oil feature information set; a cross entropy loss function is added, an improved sand dune cat swarm algorithm is adopted to optimize parameters in CatBoost, an illegal cooking oil CatBoost model is constructed, and the illegal cooking oil CatBoost model is adopted to classify and recognize the extracted illegal cooking oil feature information set; and outputting the identification result of the illegal cooking oil. The illegal cooking oil CatBoost model is adopted for illegal cooking oil identification, and the accuracy of illegal cooking oil type identification can be improved.

An image processing method, device, equipment and storage medium are provided in the embodiments of the present application. The method comprises: collecting a sequence of exposure images under a target exposure condition, the sequence of exposure images comprising a plurality of exposure images; determining a reference exposure image from the plurality of exposure images; calculating an image difference parameter between a target exposure image and the reference exposure image in the sequence of exposure images, the target exposure image being an exposure image other than the reference exposure image in the sequence of exposure images; determining a plurality of fusion exposure images in the target exposure image according to the image difference parameter; and obtaining a denoising image according to the plurality of fusion images. The image processing method provided in the present scheme reduces the workload of human participation when obtaining the denoising image, and improves the efficiency of obtaining the denoising image. On the other hand, the quality of the denoising image is also effectively improved, the fitting degree of the denoising image and the real scene is improved, and thus the effect of network denoising is improved.

The present disclosure provides a keyword recognition model training method and device. The method comprises: obtaining a training sample; the training sample comprises: a dish name of an initial dish, dish multi-domain information, and a plurality of initial keywords associated with the initial dish; inputting the dish name, the dish multi-domain information, and the plurality of initial keywords into a to-be-trained keyword recognition model; calling the to-be-trained keyword recognition model to process the dish name, the dish multi-domain information, and the plurality of initial keywords, and obtaining a prediction probability of the plurality of initial keywords output by the to-be-trained keyword recognition model with respect to the initial dish; calculating a loss function of the to-be-trained keyword recognition model according to the prediction probability; adjusting model parameters of the to-be-trained keyword recognition model according to the loss function; iteratively performing the steps of obtaining the training sample and adjusting the model parameters of the to-be-trained keyword recognition model according to the loss function until a target keyword recognition model is obtained. The present disclosure can improve the accuracy of keyword prediction.

The application provides a time-frequency feature extraction method and device for an engine vibration signal. The method comprises the following steps: dividing the vibration acceleration signal of the engine into a noise sub-signal and a vibration sub-signal, filtering the noise sub-signal by using a non-local mean method, and recombining the filtered signal and the vibration sub-signal in the time domain, processing the recombined signal by using a TVF-EMD method, further suppressing the noise signal by constructing a plurality of different filters, and obtaining a vibration acceleration signal composed of a plurality of intrinsic mode function sub-signals. The application removes the noise interference information of the vibration signal from two directions of time and frequency domains, strengthens the denoising effect, and realizes accurate extraction of vibration characteristic information.

A quantum data processing system and method are provided. A computer-implemented method may include: storing multiple copies of a quantum state in a quantum memory, wherein the quantum state encodes properties of a target system; loading the multiple copies of the quantum state in the quantum memory into a quantum computer; processing the multiple copies of the quantum state by the quantum computer to obtain a purified quantum state; and measuring the purified quantum state to determine properties of the target system.

The application relates to the technical field of signal processing, and specifically provides a magnetocardiogram denoising method, a storage medium, an electronic device and a product. The method can comprise the following steps: acquiring a training data set containing a noisy magnetocardiogram signal sample and a denoised magnetocardiogram signal sample; processing the noisy magnetocardiogram signal sample by using a multistage gating network model to obtain a predicted denoised magnetocardiogram signal; wherein the multistage gating network model comprises an encoder, a decoder and a feature fusion module; and after the multistage gating network model is optimized by using a signal loss value between the predicted magnetocardiogram signal and the denoised magnetocardiogram signal sample, a target denoising network model is obtained; wherein the target denoising network model is used for denoising a real noisy magnetocardiogram signal to obtain a denoised magnetocardiogram signal. Some embodiments of the application can realize high-quality denoising of signals and have high practicability.

This invention discloses an infrared spatiotemporal noise generation method based on hybrid neural representations. The method includes establishing an infrared denoising dataset containing paired indoor noise-clear video and unpaired outdoor noise-clear video; constructing an infrared spatiotemporal noise model based on hybrid neural representations and its training loss function; and building a spatiotemporal noise generator G and a spatiotemporal discriminator. This invention employs a divide-and-conquer approach, independently exploring the noise synthesis path from both spatial and temporal dimensions. By constructing a hybrid neural representation of noise, it deeply integrates the spatial and temporal embeddings of noise and implicitly models the complex spatiotemporal distribution of infrared noise through recurrent adversarial learning. This comprehensively and deeply characterizes the spatiotemporal properties of noise, providing new ideas and tools for noise modeling and analysis in the field of video processing.

The application discloses a human body breathing information extraction method, device and equipment based on radar and a medium. The method comprises the following steps: mixing and filtering echo signals and transmission signals to obtain original data, wherein the transmission signals are radio frequency signals transmitted by an X-band radar, and the echo signals are signals received by an antenna; processing the original data, and determining whether there is a stationary human body according to the processed data; performing phase correction on the original data with the stationary human body, and extracting phase signals; performing phase unwrapping on the extracted phase signals; extracting a chest wall displacement-time signal from the unwrapped phase signals, and obtaining a breathing frequency of a human body target. Compared with the prior art, the application has a better anti-interference effect in extracting breathing information from echo signals of an X-band radar.

The application discloses a color noise suppression method, a color noise suppression device, an electronic device and a computer readable storage medium. The method comprises the following steps: counting a color histogram of a to-be-processed image based on a preset first color dimension and a second color dimension, wherein the horizontal axis of the color histogram is the first color dimension, the vertical axis of the color histogram is the second color dimension, and the value of each coordinate point in the color histogram is used to represent the number of pixels; finding the main color and the trace color of the to-be-processed image through the color histogram; establishing a color mapping relationship of the to-be-processed image according to the main color, the trace color and the color histogram; and adjusting the to-be-processed image according to the color mapping relationship to obtain a color noise-removed image. The application can balance the denoising effect and the processing complexity, and help to quickly obtain an image with low color noise.

This invention provides a method for extracting the fluorescence boundary line of laser-induced fluorescence emitted by transparent ice. The method involves brightness segmentation of the image, dividing it into brighter and darker regions. Based on the laser's position, the boundary region between the bright and dark areas near the laser is selected as the coarse localization region for the fluorescence boundary line. Within this coarse localization region, the image gradient is calculated, and the direction y of the bright-dark boundary line is calculated based on the gradient. Based on the direction y, an m*d sampling area is set, and within this sampling area, the image boundary E is calculated. This method effectively extracts the coordinates of the laser-induced fluorescence boundary line. Based on the system calibration results, the contour line of the transparent ice body can be measured. Combined with a 3D scanning device, the 3D morphology of the ice can be measured through 3D point cloud stitching, solving the problem that existing line structured light extraction methods cannot be used for 3D morphology measurement of laser-induced fluorescence transparent ice bodies.

The application discloses a NSST domain sonar image denoising method based on density clustering and gray scale transformation, and belongs to the technical field of sonar image denoising.The application solves the problems of poor denoising effect, poor edge feature maintaining capability and low denoising efficiency of the existing sonar image denoising method.The application can improve the low contrast and serious noise interference of a sonar image, decomposes the noisy sonar image through NSST transformation, removes noise signals in high-frequency coefficients by using density clustering, retains detail signals, performs gray scale transformation on low-frequency coefficients, improves image contrast, and finally performs NSST inverse transformation to obtain a denoised sonar image.The method can be applied to denoising of a sonar image.

The application belongs to the field of magnetic resonance sounding signal noise filtering, and is a kind of magnetic resonance sounding signal extraction method based on intelligent optimization manifold learning, the parameter group of the local linear embedding manifold learning method is initialized, the genetic algorithm in the intelligent optimization algorithm is used, the signal-to-noise ratio is taken as the fitness function, and the parameter group in the local linear embedding manifold learning method is optimized, the local linear embedding manifold learning method uses the optimized parameter group to sequentially perform first processing and second processing on the magnetic resonance sounding signal, removes random noise, and obtains the final denoised magnetic resonance sounding signal. The application effectively retains the signal characteristics through the nonlinear dimension reduction of manifold learning, avoids the information loss caused by frequency band selection, maintains the local relationship between data points in the dimension reduction process, ensures that adjacent points in m-dimensional space remain adjacent in d-dimensional space, only compresses and filters noise, and realizes signal extraction.

The invention relates to an intelligent pressure bra and a control method thereof. Comprising a bra main body, a pressure sensing module, a posture sensing module, a pressure adjusting module and a central control module, the pressure sensing module, the posture sensing module and the pressure adjusting module are fixedly connected to the bra main body; the pressure sensing module is used for collecting original chest pressure signals of a patient; the posture sensing module is used for collecting an original posture signal of a patient; the pressure adjusting module and the pressure sensing module are correspondingly arranged; the central control module is connected with the pressure sensing module, the posture sensing module and the pressure adjusting module and used for processing data, communicating with remote terminal equipment and controlling the pressure adjusting module. Compared with the prior art, the method has the advantages that real-time closed-loop adjustment is realized; the compression strategy is dynamically optimized in combination with the posture of the patient, and the safety and accuracy of postoperative rehabilitation are improved.

This invention relates to a speech denoising method and device based on deep compressed sensing, comprising the following steps: adding Gaussian white noise to a clean speech signal to obtain a noisy speech signal; performing time normalization and frame segmentation on both the clean and noisy speech signals; extracting perceptual features from both the clean and noisy speech signals to obtain perceptual features of the clean speech and the noisy speech; training the GAN using the perceptual features of the noisy speech as input to a generative adversarial network (GAN) and the perceptual features of the clean speech as the objective function to obtain perceptual features for generating clean speech; and reconstructing the generated clean speech signal using the OMP reconstruction algorithm based on compressed sensing. This invention, through deep compressed sensing, is suitable for nonlinear noise analysis and processing, improving the denoising effect of noisy speech signals.

Embodiments of the present application provide a method, device, equipment and medium for generating a labeled ultrasound sample image. In the method, a diffusion model and an auxiliary segmentation network are obtained by pre-training. A target mask image is obtained, and a guide condition is determined based on the target mask image and the auxiliary segmentation network. A target sample image corresponding to the target mask image is generated by the diffusion model based on the guide condition. The target mask image is used to represent a target object contained in the target sample image. In this way, the auxiliary segmentation network and the target mask image can be used as guide conditions to provide directional guidance for the diffusion model, thereby improving the denoising effect of the diffusion model and improving the sample quality of the target sample image generated by the diffusion model, while reducing the difficulty and cost of obtaining the target sample image.

The invention discloses a low-illumination image enhancement method based on bidirectional cross-frequency-domain guided wavelet diffusion, belongs to the technical field of computer vision, and solves the problems that an existing diffusion model is high in calculation cost, easy to amplify noise and difficult to balance global and local details in low-illumination enhancement. Each of the training set and the test set comprises a low-illumination image and a reference image; decomposing the image into low-frequency and high-frequency components through discrete wavelet transform; only recovering a low-frequency component by using a diffusion model, introducing a detail weighted loss mechanism in training, and guiding low-frequency recovery by using high frequency; using the recovered low-frequency component as a priori, combining with a noisy high-frequency component, and performing feature modulation and denoising through a cross-frequency-domain gating refinement network; reconstructing a final enhanced image based on inverse wavelet transform; constructing a bidirectional cross-frequency-domain guided wavelet diffusion network model and designing a multi-objective loss function for training; and for the low-illumination image in the test set, the operation is executed based on the trained network model, and a high-fidelity enhanced image is obtained.

This invention provides a low-dose CT image reconstruction method, belonging to the field of medical image processing technology. The method is characterized by the following steps: Step A, obtaining two different low-dose, high-noise CT image data, one as input image I and the other as a self-supervised image T; Step B, randomly dividing input image I into proportionally sized image blocks; Step C, randomly removing 60% of the proportionally sized image blocks from Step B, randomly sorting the remaining 40% of the image blocks, and inputting the sorted image block sequence position information into the algorithm model; Step D, after reconstruction by the algorithm model, obtaining the reconstructed image pairs and calculating the L1 loss with T. This reconstruction method requires convenient image data acquisition, allows for active adjustment of the image noise reduction ratio, and the reconstructed image does not reduce the image signal-to-noise ratio, nor weaken details and textures, resulting in high image quality.

The invention discloses a digital holographic microscopic imaging coherent noise suppression network model and method based on deep learning. The network model adopts a double-branch encoder-single decoder structure, and a double-branch encoder extracts local details and global noise features of an image in parallel; the innovative double-branch intensive attention fusion module fuses multi-scale features through an adaptive weighting and enhancement mechanism; and the decoder integrates information through jump connection and reconstructs a clear image. The training method generates a high-fidelity data set based on a physical simulation system. According to the method, speckle noise and parasitic fringes can be efficiently suppressed, the phase details of the object are effectively reserved while the peak signal-to-noise ratio and the structural similarity index are remarkably improved, and the method has excellent generalization ability and calculation efficiency and is suitable for real-time high-quality imaging of a digital holographic microscopy system.

This invention provides a compressed spectral imaging reconstruction method and apparatus based on a degradation estimation recurrent neural network. By sharing parameters across different stages, the deep unfolded network is converted into a recurrent neural network. This parameter sharing significantly reduces the number of network parameters and memory usage, and allows the deep unfolded network to learn representations for reconstruction from inputs at different stages. This reduces the parameter sparsity of the deep unfolded network, fully exploits its reconstruction potential, and enhances its reconstruction performance, thereby improving the quality of the reconstructed image. Furthermore, this invention introduces a degradation estimation network to estimate the degradation matrix and noise level, improving the accuracy of solutions to the data and prior subproblems. Simultaneously, the use of local and non-local transformation networks to solve the prior subproblems more fully exploits the local and non-local priors of the hyperspectral image, further improving the quality of the reconstructed image.

The application discloses a double-end traveling wave fault location method based on wavelet transform and Teager-Kaiser energy operator (TKEO), which is applied to a 800V direct current hydrogen production line. In order to improve the accuracy of fault identification and the precision of fault location, firstly, wavelet transform is performed on a fault signal, a soft threshold is used for denoising and the signal is reconstructed; characteristic information of each wavelet decomposition layer is extracted and analyzed, and the fault type is determined according to the energy ratio of high-frequency and low-frequency decomposition layers; secondly, TKEO is used to extract the instantaneous energy spectrum after wavelet decomposition, and the sampling points of the first wave head reaching the two ends of the direct current line are accurately calibrated; finally, the double-end ranging method is used to accurately solve the fault distance.

The invention discloses an improved BM3D image denoising method based on multi-element fusion block similarity measurement. According to the method, wavelet decomposition is carried out on a noise image, a noise standard deviation is stably estimated on a highest-frequency sub-band, BM3D parameters are initialized, in a basic estimation stage, pixel intensity, local variance and gradient difference are fused to construct block similarity measurement, similar block matching grouping and three-dimensional block group construction are realized, and the method has the advantages of being high in robustness and high in robustness. And in a final estimation stage, a basic estimation image is obtained through three-dimensional collaborative hard threshold filtering and weighted aggregation, in the final estimation stage, secondary matching is guided by the basic estimation image, a three-dimensional block group is constructed again, a structure and texture adjustment factor is introduced to form an adaptive Wiener coefficient for Wiener collaborative filtering, and finally, a target de-noised image is output through weighted aggregation. In the whole process, the phenomena of mismatching and over-smoothing of a traditional BM3D algorithm under the condition of high noise can be reduced, the edge and texture detail keeping capability is enhanced, and the denoising quality of a noise image is improved.

The invention discloses a hardware implementation method of a time-space domain filtering system combined with image information. The hardware implementation method comprises the following steps: preprocessing an image by adopting downsampling layering and upsampling phase difference; processing the data stream of each layer of original image preprocessing graph in a pipeline form to obtain four-direction edge filtering image video stream data; performing read-write processing on each layer of filtered image video stream data to obtain multi-frame image cache data; performing motion compensation denoising on each layer of filtered image in a time domain; performing kernel convolution on the data video stream after motion compensation denoising through a bilinear interpolation formula in a pipeline form to obtain image data recovered to an original resolution, and fusing video stream data of each layer; and processing the fused image data in an assembly line form to obtain image video stream data after spatial domain processing. According to the method, the detail information in the image is reserved as much as possible while the image noise is suppressed, and the moving target trailing phenomenon in the image scene is effectively suppressed under the condition of the moving scene.

The invention relates to the technical field of video processing, and discloses an edge computing and AI fused real-time video stream processing optimization method. The edge calculation and AI fused real-time video stream processing optimization method comprises the following steps: preprocessing a real-time video; performing video optimization processing; obtaining an optimization result; according to the invention, the internal noise of the video is filtered by using a block and low-rank tensor recovery video denoising algorithm, so that the video definition is improved, and the video edge information is more significant; the edge of a denoised video is detected through a PSVM, the video edge detection effect is improved, the noise of an original video image can be effectively removed through an algorithm, the definition of the video image is improved, edge information is made to be more remarkable, when the noise combination level is higher, the entropy value, the AG value and the PSNR value of the two video sequences which are expanded and denoised through the method are all in a descending trend, and the noise combination level is higher. Therefore, the method provided by the invention has a better denoising effect when the video sequence with a simple and complex structure is processed.