66results about How to "Improve recovery quality" patented technology

The invention provides a Bayer color image interpolation method based on direction flag bits. The method includes the steps that firstly, unknown green components are restored, gradients of each pixel point in four directions are detected through edge detection operators, direction flag bit information of each pixel point is determined according to the detected gradients, the chromatic aberration coefficient of each pixel point and an adjacent pixel point is calculated, and interpolation restoration is conducted on the unknown green components according to the direction flag bit information and the chromatic aberration coefficient. After all the green components of a whole image are restored, all unknown red and blue components are finally restored based on a chromatic aberration method, and then color interpolation conducted on the Bayer image is finished finally. Compared with a traditional method that single-channel independent interpolation is conducted only by detecting horizontal and perpendicular edges, on the basis that the low complexity and the high image quality of an existing method are kept, interpolation accuracy is further improved, and the method is more suitable for human eye observation and provides a foundation for high-accuracy image restoration in the following process.

A method for restoring hyperspectral remote sensing image based on non-convex and low-rank sparse constraint belongs to the field of hyperspectral remote sensing image processing in remote sensing image processing. In order to solve the problem that the existing hyperspectral remote sensing image restoration technology can not effectively remove noise and improve the image restoration quality, themethod comprises the following steps: inputting a hyperspectral remote sensing image; initializing a weight coefficient matrix, iterative times and a convergence threshold, initializing sub-image size and scanning step, partitioning sub-blocks; establishing an image restoration model; the auxiliary variable and the coefficient of the regular term being introduced, and the maximum-minimum algorithm being used to solve the problem iteratively; judging whether the restoration result satisfies the convergence condition; obtaining a hyperspectral restored image that meets the requirements by iterative times, otherwise returning to corresponding steps to continue the iterative operation; calculating a weight coefficient matrix and assigning appropriate weights to each sub-block; hyperspectral remote sensing images being restored to obtain the final restored hyperspectral remote sensing images. The effect of denoising is obvious and the image details are preserved.

A transient signal encoding method and device, decoding method and device, and processing system, where the transient signal encoding method includes: obtaining a reference sub-frame where a maximal time envelope having a maximal amplitude value is located from time envelopes of all sub-frames of an input transient signal; adjusting an amplitude value of the time envelope of each sub-frame before the reference sub-frame in such a way that a first difference is greater than a preset first threshold, in which the first difference is a difference between the amplitude value of the time envelope of each sub-frame before the reference sub-frame and the amplitude value of the maximal time envelope; and writing the adjusted time envelope into bitstream.

A transient signal encoding method and device, decoding method and device, and processing system, where the transient signal encoding method includes: obtaining a reference sub-frame where a maximal time envelope having a maximal amplitude value is located from time envelopes of all sub-frames of an input transient signal; adjusting an amplitude value of the time envelope of each sub-frame before the reference sub-frame in such a way that a first difference is greater than a preset first threshold, in which the first difference is a difference between the amplitude value of the time envelope of each sub-frame before the reference sub-frame and the amplitude value of the maximal time envelope; and writing the adjusted time envelope into bitstream.

The invention discloses a method for generating and authenticating digital image tampered content recoverable variable capacity watermarks, which comprises the following steps of: dividing 2*2 image blocks into smooth image blocks and non-smooth image blocks, extracting 6bit characteristics from the smooth image blocks, extracting 12bit characteristics from the non-smooth image blocks, encrypting the characteristics of the image blocks to generate image block recovery watermarks, randomly embedding the image block recovery watermarks into other image blocks, and judging the authenticity of the image blocks by comparing the consistency of the characteristics of the image blocks and combining block neighborhood characteristics. For the judged tampered image blocks, different tampering recovery operations are executed in two steps according to whether the corresponding recovery watermarks of the image blocks are tampered to improve the tampering recovery quality. The variable capacity recovery watermarks are generated according to the personal characteristics of the image blocks, and the recovery watermarks are simultaneously used for tampering positioning and tampering recovery, so the watermark embedding capacity is reduced when enough image characteristics are ensured, and watermark-containing image quality and tampering recovery quality are considered; and meanwhile, the security is improved, and known forging attacks such as joint attack, mean attack and the like can be resisted.

An image processing method includes (a) performing edge detection to an image to obtain a plurality of the edge pixels of the image; (b) performing partial PSF estimation to each of the edge pixels to generate a plurality of partial PSF estimation results; and (c) generating a PSF estimation result according to the partial PSF estimation results.

Disclosed is a low-complex adaptive block compressing sensing image coding method based on visual perception. The low-complex adaptive block compressing sensing image coding method based on visual perception is characterized in that I. Block compressing sensing: 1. Coding: reading images and separating images; block observing; sending observed values and observed rates to decoding end; 2. Decoding: measurement matrix is obtained from the observed rates; finding an initial solution; wiener filtering; updated by procedural language (PL); discrete cosine transform (DCT) switching; bivariate threshold denoising disposing; inverse discrete cosine transform (IDCT); updated by the PL; repeated till finish of initial phrase; II. Visual analysis: analyzing each refactoring DCT coefficient block; classifying blocks; entering feedback channel; III. The adaptive block compressing sensing instructed by visual perception: 1. Coding: adaptive observe according to feedback results;sending observed values and observed rates to decoding end; 2. Decoding: combining the observed values; last phrase recover image as a first value; wiener filtering; updated by PL; DCT switching; bivariate threshold denoising disposing; IDCT; updated by the PL; till end of the decoding; next phrase analysis, adaptive observe and refactoring, till the recover image meets the need.

The invention discloses a self-adaptive combination filtering method under an impact noise condition. According to researches, an impact noise has an approximate sparse characteristic, and according to the approximate sparse characteristic, the impact noise has large amplitude in limited time, and has small amplitude in other time. A group sparse characteristic is shown, and according to the group sparse characteristic, most of sample values of signals in a time domain are zero, and at the same time, nonzero sample values occur in a grouped / clustered way. By using the characteristic of the impact noise and combining with the own characteristics of the signals, a target function is reconstructed, and a signal / noise combination estimation algorithm is designed to improve signal recovery quality. By adopting algorithm theoretical analysis and computer simulated analysis, the provided algorithm performance is more excellent, and a good application prospect is provided in a parameter estimation field, a voice signal processing field, and other fields.

Provided is a binary image encryption method based on diaphragm encryption and a phase retrieval algorithm. After a binary image and a diaphragm which are to be encrypted are encrypted through a double-random-phase encryption system in a Fresnel domain, two strength distributions are obtained, and the two strength distributions are added according to a certain proportion to obtain a ciphertext; a holographic technique does not need to be utilized to record phase information in an optical encryption process, a decryption process is achieved through the phase retrieval algorithm, and the diaphragm serves a secrete key in the decryption process. The binary image encryption method based on diaphragm encryption and the phase retrieval algorithm has the advantages that the design of a light path is simple, the decryption operation is high in convergence rate and system safety is high.

The invention discloses a real image noise reduction method based on low-rank approximation, and belongs to the field of computational photography. The implementation method comprises the following steps: establishing an RGB real image noise reduction model, carrying out block matching operation, searching a similar block for each local block, and forming a data matrix; calculating a weight matrixused for balancing real noise; performing low-rank approximation on each formed data matrix to obtain a denoised data matrix; and recovering the data matrix into blocks, and finally integrating the blocks into a noise-reduced recovered image, namely generating a high-quality real image noise-reduced recovered image. The noise reduction work of the real image can be completed with high quality, and the recovery effect of the real noise image is improved after the noise characteristics of the real image are fully considered. The method is suitable for multiple fields of medicine, military affairs, agriculture and the like.

The invention discloses a three-dimensional image semi-fragile watermarking method based on binocular just-perceived distortion, comprising the following steps of: firstly, extracting positioning watermarks of left and right images; then, determining embedded bit positions of the positioning watermarks on the left image according to a binocular just-perceived distortion value of the left image, so as to embed the positioning watermarks of the left and right images into the left image by layers; and finally, utilizing discrete cosine transform and JPEG (Joint Photographic Experts Group) quantitative compression to embed right image recovering information into the right image; and with regard to the left image, embedding recovering information after an exposure shielding region is compressed into the left image. The three-dimensional image semi-fragile watermarking method has the advantages that the method sufficiently utilizes the matching property of the left and right images so that the quality of a distorted three-dimensional image is further improved; and an experimental result shows that the method not only can accurately position a distorted region, but also can recover the content of the distorted region according to a relation of the left and right images, and the recovering information.

The invention discloses a method restraining image impulse noise based on an iteration non-local mean value. The method is mainly used for solving the problems that an existing method can not retain image detail information in the denoising process, and a sharp image can not be restored under high noise density. The method comprises the following realizing steps that (1) for a noise image, a histogram method is used for detecting the noise position; (2) an option switch median filter is used for carrying out pre-filtering on the noise image; (3) the iteration non-local mean value method is used for the image processed in a pre-filtering mode is carried out so that the sharp image with the impulse noise effectively restrained can be obtained. A simulation experiment shows that under the circumstance of the impulse noise of different degrees, the method is better in subjective visual effect and objective evaluation result compared with an existing algorithm, and the method can be used for restraining the high-density impulse noise and restoring the sharp image.

The invention discloses a cloud-containing remote sensing image compression method based on a filling strategy, and mainly solves the problems that in the prior art, the edge of a filled cloud regionstill has more high-frequency information, and the coding consumption is higher. According to the scheme, the method comprises the steps of carrying out boundary extension on an input cloud-containingimage and an input cloud mask image; calculating an approximate median of a region of interest in the cloud-containing image; obtaining a gray mask of a marked cloud region and a gray image of the edge of the marked cloud region; calculating a pixel mean value in the edge of each cloud region and filling the corresponding cloud region; carrying out edge mean filtering on the filled image; performing level shifting on the filtered image by using the approximate median of the region of interest to obtain a preprocessed image; encoding the preprocessed image to obtain a compressed code stream, and sending the compressed code stream to a receiving end; and the receiving end decoding the received code stream to obtain a restored image. The method can effectively reduce the coding rate of the cloud region, improves the recovery quality of the region of interest, and can be used for the transmission of various cloud-containing images.

The present invention provides a method for erasing a visible watermark of a document image. The method comprises: carrying out binaryzation preprocessing on a target image to obtain a feature image; determining watermark areas and non-watermark areas of the target image and the feature image; respectively partitioning the non-watermark areas of the target image and the feature image into a plurality of overlapped blocks and carrying out sampling on each block; constructing a sparse representation model; optimizing the constructed sparse representation model by utilizing a dictionary learning method so as to obtain a compressed target image block dictionary and a feature image block dictionary; and according to the similarity matching degree of the two dictionaries, erasing the visible watermark in the target image by adopting a dictionary similar block erasing algorithm or a local erasing algorithm in a self-adaptive mode. According to the method, the visible watermark of the document image can be effectively erased; moreover, context semantic readability of background texts in an original watermark area can be kept; and the original background texts can be recovered to the greatest degree.

The embodiment of the invention provides an image recovery method and device based on a pre-training auto-encoder. The method comprises the following steps: obtaining a speckle image; inputting the speckle image into an image restoration network model to obtain a restored image, wherein the image recovery network model comprises a first encoder sub-network and a first decoder sub-network which aretrained, the initial parameters of the first decoder sub-network are determined according to parameters in a pre-trained auto-encoder network model, and the auto-encoder network model is obtained bytraining according to a first training set, the image restoration network model is obtained by training according to a second training set, and the second training set comprises a second sample original image and a sample speckle image. According to the method, the auto-encoder network model is pre-trained, and the parameters in the trained auto-encoder network model are adopted to perform parameter initialization on the first decoder sub-network in the image recovery network model, so that training data set information can be fully utilized, and the convergence rate of the network during training is increased.

The invention discloses a thumb rehabilitation training device with multiple degrees of freedom for hemiplegic patients. The device includes a supporting base, a driving sensing mechanism I and a driving sensing mechanism II, wherein the driving sensing mechanism I and the driving sensing mechanism II are fixed on the supporting base, and composed of optical-electricity encoders, direct-current torque motors, magnetorheological fluid dampers and torque sensors; the output end of the driving sensing mechanism I is connected to one end of an arc-shaped guiding rail through a coupler I, and the other end of the arch-shaped guiding rail is fixed on a supporting plate I successively through a connecting rod I, a connecting rod II and a connecting rod III; the connecting rod III is fixed on the supporting plate I through a bearing I, and the supporting plate I is fixed on the supporting base; the output end of the driving sensing mechanism II is connected to a driving shaft of a finger training mechanism through a coupler II, the finger training mechanism includes a training rod, a front finger sleeve, a rear finger sleeve, a driving connector and the driving shaft, and the driving shaft is connected with a training rod in a transmission mode through the driving connector; the front end of the training rod is embedded into the arc-shaped guiding rail and slides along the arc-shaped guiding rail.

The invention provides a preparation method of an exosome as well as the exosome prepared by adopting the method and applications of the exosome. The method comprises the following steps: (1) collecting the human adipose tissue, and carrying out extracting to obtain adipose-derived stem cells; and (2) carrying out treatment on the adipose-derived stem cells obtained in step (1), and thus the exosome is obtained through extraction. The exosome is the exosome derived from the adipose-derived stem cells, when the exosome is used for skin repair, the wound healing speed can be accelerated, the pain of patients can be alleviated, the wound recovery quality can be improved, and the area of scars can be reduced. The average time for the wound healing of the untreated group is 33.7 days, after thegel preparation containing the exosome provided by the invention is applied clinically, the wound healing only needs 12.2 days, thus the pain of the patients of is obviously alleviated, and the woundinfection probability is reduced; the effect for the refractory wound such as diabetes is also very obvious, the effective rate achieves 85.7%, but the effective rate of the control group is only 42.1%.

The invention discloses an audio bandwidth expansion method and an expansion apparatus thereof. The method comprises the following steps of detecting a coding mode of a current frame signal in a mixing ACELP / TVC core encoder so as to distinguish a signal type; based on the signal type, selecting adaptive high frequency reconstruction strategies for voice and music signals respectively; if the signal is the voice signal, adopting a bandwidth expansion method based on LPC; and if the signal is the music signal, adopting a frequency band copy bandwidth expansion method based on a high and low frequency signal correlation. The expansion apparatus comprises a signal type detection module, a voice signal bandwidth expansion module and a music signal bandwidth expansion module. In the invention, characteristics of different signal types are fully considered; and from an aspect of the signal type, audio recovery quality is increased and high frequency reconstruction can be accurately performed.

The invention discloses a three-dimensional image fragile watermark method based on joint just noticeable distortion and in particular relates to a three-dimensional image location tampering and restoring fragile watermark algorithm based on JJND (Joint Just Noticeable distortion). Firstly, a three-dimensional image block is divided into a relative sensitive block and a relative insensitive block according to a JJND model and the number of three-dimensional image match blocks, then discrete cosine transform and quantization compression are carried out on the image blocks to generate watermarks, and unsymmetrical watermark embedding of different type blocks of the three-dimensional image is built according to the JJND model and types of the image blocks. Restored watermarks can be used for locating the positions of the tampered portions to further improve transparency of the image. Meanwhile, copy embedding of the restored watermarks in the three-dimensional image can improve the capacity for the watermarks to restore tampered information. Experiments show that the three-dimensional image fragile watermark method based on the joint just noticeable distortion is accurate in location on the tampered positions and has higher restoring capacity on the tampered areas.

The invention discloses a single-carrier optical communication signal detection method and a single-carrier optical communication signal detection device. A first optical branching module divides an optical signal into a receiving signal L1 and a receiving signal L2; a time delay module performs time delay processing on the L1 to obtain a time delay optical signal; a second optical branching module divides the time-delay optical signal into a time-delay optical signal M1 and a time-delay optical signal M2; the direct detection module directly detects the M1 to obtain an electric signal R1; thedifferential detection module performs differential detection on the L2 and the M2 to obtain an integral electric signal R2; the first calculation module performs subtraction on the electric signal R1 and the electric signal R2 to obtain a detection electric signal R3; the conversion module preprocesses the detection electric signal R3 and performs analog-to-digital conversion to obtain a digitalelectric signal R; the digital signal circuit module processes the R through a digital signal circuit and then outputs a sending end digital baseband signal. According to the invention, the problem of low sensitivity caused by SSBI interference in the existing direct detection is solved, and the complex-valued signal direct detection device with higher transmission sensitivity and higher spectrumutilization rate is obtained under the conditions of low complexity and low cost.

The embodiment of the invention provides an image processing method and system and a storage medium. The method comprises the following steps: inputting a pure phase mask into a spatial light modulator, so that the spatial light modulator modulates a laser beam incident to the spatial light modulator based on the pure phase mask; performing bit plane layering processing on the target image to obtain a plurality of binary images; a plurality of binary images with preset spacing distances are arranged between a spatial light modulator and a barrel detector, and light beams emitted by the spatial light modulator penetrate through the binary images and then reach the barrel detector, so that the barrel detector obtains light intensity based on the position information of the pure phase mask and the binary images; and arranging all the light intensities corresponding to one binary image into one-dimensional intensity vectors, and taking a plurality of one-dimensional intensity vectors of which the number is equal to that of the binary images as an encryption result of the target image. The image restoration quality can be improved.

The invention discloses a natural image super-resolution method based on an expectation maximization algorithm, which comprises the steps of (1) inputting a low-resolution mage; (2) carrying out interpolation on the image; (3) acquiring a hidden image; (4) cutting the hidden image into hidden image blocks; (5) acquiring a similar matrix of each hidden image block; (6) acquiring a dictionary of each hidden image block; (7) acquiring the mean and the covariance of each estimation image block; (8) acquiring the maximum posterior estimation value of each estimation image block; (9) acquiring a high-resolution image; (10) calculating a relative error; (11) judging whether the relative error meets a termination condition or not; (12) updating data; and (13) outputting an optimal high-resolution image. According to the invention, the expectation maximization algorithm is introduced into the field of natural image super-resolution, and abundant information for recovering image details is acquired, thereby being suitable for image super-resolution under complex conditions.

The invention discloses a traffic image deblurring method and device based on multi-scale adversarial learning. The method comprises the following steps of: 1, generating a clear-fuzzy traffic image pair by adopting a motion blur generation method; 2, constructing a traffic image deblurring network model based on adversarial learning, and carrying out adversarial learning on the recovered image to enable the image to be closer to a real clear image; Step 3, constructing a plurality of image deblurring network models with the same structure and the scales from small to large, transmitting a clear image recovered under a small scale as input to a subsequent network, recovering scale by scale, and finally obtaining a deblurring image of the original size; And step 4, carrying out joint training on the multi-scale adversarial learning image deblurring network model. According to the method, the quality of the deblurred image can be improved through a multi-scale adversarial learning method.

An image processing method includes (a) performing edge detection to an image to obtain a plurality of the edge pixels of the image; (b) performing partial PSF estimation to each of the edge pixels to generate a plurality of partial PSF estimation results; and (c) generating a PSF estimation result according to the partial PSF estimation results.

The invention discloses an iterative depth map structure restoration method based on RGB-D SSIM structure similarity. First, the edge of the input depth map is detected, and the edge is expanded, and the expanded area is marked as a potential structural distortion area, and then the Each pixel in the potential structural distortion area judges whether it is distorted, and generates a structural distortion metric index. The distorted pixel uses the product of the Gaussian weight of the color image and the structural distortion metric index to construct a restoration weight, and performs guided restoration through a weighted median filter, and then The distorted area is guided and filtered, and the completed result map continues to iterate according to the above steps until the set termination iteration condition is met, and the depth map is output to end the calculation. The present invention obtains more accurate structural information by iteratively detecting and recovering the structurally distorted area of ​​the depth map, and at the same time denoises and preserves the edges of the structurally undistorted area, finally obtaining a depth map with a clear structure and smooth depth values .

The invention discloses a natural image super-resolution method based on an expectation maximization algorithm, which comprises the steps of (1) inputting a low-resolution mage; (2) carrying out interpolation on the image; (3) acquiring a hidden image; (4) cutting the hidden image into hidden image blocks; (5) acquiring a similar matrix of each hidden image block; (6) acquiring a dictionary of each hidden image block; (7) acquiring the mean and the covariance of each estimation image block; (8) acquiring the maximum posterior estimation value of each estimation image block; (9) acquiring a high-resolution image; (10) calculating a relative error; (11) judging whether the relative error meets a termination condition or not; (12) updating data; and (13) outputting an optimal high-resolution image. According to the invention, the expectation maximization algorithm is introduced into the field of natural image super-resolution, and abundant information for recovering image details is acquired, thereby being suitable for image super-resolution under complex conditions.

The invention discloses a method for suppressing the boundary ringing effect of image restoration based on sinusoidal integration. First, the blurred image f(x,y) is input, the blur kernel h(x,y) is calculated, and then the original blurred image is mirrored. Extend to obtain the extended image g(x,y); then open up the area of ​​the window function w(x,y) according to the size of the extended image g(x,y), and use the sine function integration method to fill the window function; then The extended image g(x,y) and the window function w(x,y) are synthesized to obtain the windowed image g'(x,y); finally, the image restoration algorithm is used to perform the windowed image g'(x,y) Restore and crop out the original image size of M×N in the middle area, which is the final restored image. The present invention suppresses the ringing effect in the restoration of motion blurred images, not only does not extend the image too much to increase the amount of calculation, but also does not ignore the detailed information of the image edge, thereby improving the restoration quality of the blurred image.

The invention discloses a method for generating and authenticating digital image tampered content recoverable variable capacity watermarks, which comprises the following steps of: dividing 2*2 image blocks into smooth image blocks and non-smooth image blocks, extracting 6bit characteristics from the smooth image blocks, extracting 12bit characteristics from the non-smooth image blocks, encryptingthe characteristics of the image blocks to generate image block recovery watermarks, randomly embedding the image block recovery watermarks into other image blocks, and judging the authenticity of the image blocks by comparing the consistency of the characteristics of the image blocks and combining block neighborhood characteristics. For the judged tampered image blocks, different tampering recovery operations are executed in two steps according to whether the corresponding recovery watermarks of the image blocks are tampered to improve the tampering recovery quality. The variable capacity recovery watermarks are generated according to the personal characteristics of the image blocks, and the recovery watermarks are simultaneously used for tampering positioning and tampering recovery, so the watermark embedding capacity is reduced when enough image characteristics are ensured, and watermark-containing image quality and tampering recovery quality are considered; and meanwhile, the security is improved, and known forging attacks such as joint attack, mean attack and the like can be resisted.