49results about How to "High structural similarity" patented technology

The invention discloses a cross-modal medical image registration method. The method comprises the following steps: providing a training set comprising a floating image of a first modal and a referenceimage of a second modal; inputting the floating image into an image conversion network, and converting the floating image into a conversion image of a second modal; inputting the floating image and the reference image into a cross-modal flow sub-network to output a first deformation field; inputting the converted image and the reference image into the single-mode flow sub-network to output a second deformation field; inputting the first deformation field and the second deformation field into a deformation field fusion network to output a final deformation field; inputting the floating image and the final deformation field into a space transformation network to obtain a floating image after distortion transformation of the final deformation field; obtaining a first total loss function according to the transformed floating image and the reference image, and performing supervised training on the network by taking minimization of the first total loss function as a target; and inputting ato-be-registered image into the trained network to obtain a registered image. The cross-modal medical image registration method can greatly improve the cross-modal medical image registration effect.

The invention discloses an image super-resolution method based on dictionary learning and non-local total variation. The image super-resolution method based on the dictionary learning and the non-local total variation mainly solves the problems of ring effect, lost high-frequency information, and inaccuracy of boundary matching of a super-resolution method and the like. The achieving steps include (1), inputting an image training set; (2) using a KSVD algorithm to train two corresponding high-resolution dictionary and a low-resolution dictionary; (3) conducting sparse representation of a low-resolution input image, and determining sparse coefficient; (4) using determined sparse coefficient and the high-resolution input image to obtain a high-resolution image; (5) conducting ring effect removal of the non-local total variation on the reconstructed high-resolution image; (6) conducting high-frequency information enhancing of the high-resolution image by error compensation, and obtaining a final result. By showing of a simulation experiment, compared with the prior art, the image super-resolution method based on the dictionary learning and the non-local total variation has the advantages of being simple in operation, small in noise, clear in edge and the like, and can be used for obtaining a high-definition image.

The invention provides a construction method and application of a coronavirus antigen on the basis of a mosaic strategy. The method comprises the following steps of: (1) analyzing the natural structural protein of a human coronavirus, and obtaining cytotoxic T lymphocyte epitope which is in the presence on the natural structural protein; and (2) adopting the mosaic strategy to analyze the structural protein of the human coronavirus, and after rare epitope with low occurrence frequency in the natural structural protein is removed, obtaining the coronavirus antigen assembled by short peptides. According to the construction method, on the basis of an artificial intelligence algorithm, all known CoV (coronavirus) genomic sequences are subjected to big data analysis, and a novel antigen sequence is designed on the basis of the mosaic strategy. A novel antigen covers most CTL (cytotoxic T lymphocyte) epitope of natural S, M, N and E proteins on all seven known HCoVs (human coronavirus), hasa structural similarity with a natural protein, and has a potential application prospect and value on an aspect of preparing a coronavirus universal vaccine.

The present invention provides a method of lowering intraocular pressure which comprises administering a therapeutically effective amount of a pharmaceutical composition comprising 4-bromo-5-(2-imidazolin-2-ylamino)benzimidazole, or a salt thereof to the affected eye of a patient, as a single dose, wherein the affected eye has an intraocular pressure less than the baseline intraocular pressure for at least eight (8) hours.

The invention discloses environment-friendly repeated-writing paper and a preparation method therefor. According to the environment-friendly repeated-writing paper and the preparation method therefor, a hydroxyl protective group is added into a nano bacterial cellulose suspension with excellent characteristics such as high water absorptivity and water retention ability and good water and air permeability, so as to lower hydrophilicity; a staged reaction is carried out under strong base conditions, so that bacterial cellulose reaches moderate water absorptivity and water retention ability, ink can only be present in a surface layer and be clear in writing while the ink is absorbed, and thus, writing, reading and subsequent handwriting wiping are facilitated; then, a scientifically-compounded dispersant is added, and ultrasonic and high-voltage pulse electric field combined dispersion is carried out, so that a stable and uniform modified bacterial cellulose suspension is obtained; finally, the surface of raw paper is coated with the suspension so as to obtain the writing paper with appropriate ink absorbing uniformity and adhesive power and a dense structure, thus, the writing property of the writing paper can be enhanced, and the clearness of writing is enhanced; and waterproofing and oil-proofing actions can be achieved, most importantly, ink marks on the paper can be erased with a cleanser and the paper can be reused for 40 to 100 times under the condition that the performance of the paper is not affected.

The invention discloses an OCTA imaging method and device based on two-dimensional composite registration. The invention discloses a two-dimensional composite registration method, a plurality of preliminary blood flow contrast images are used; after an OCTA effective strip is extracted, a global registration method based on large vessel skeleton registration degree and SIFT feature point extraction and a local registration method based on SIFT flow target function detection are used in a combined mode to perform registration on the OCTA effective strip pair on the two-dimensional level, and then motion artifacts are eliminated. The obtained fused OCTA result image is small in pixel average error, good in quality and high in structural similarity with an original OCTA blood flow contrast image, motion artifacts contained in the fused OCTA result image are greatly reduced and even disappear, and high-quality OCTA blood flow contrast imaging is achieved.

The invention discloses a lead-zinc flotation foam image space-time joint denoising method based on bubble motion stability analysis. The method is based on the motion characteristics of bubbles. thestable motion state (SMS) and the unstable motion state (UMS) of the foam image sub-block are accurately identified; carrying out time domain filtering denoising on the sub-blocks with the SMS characteristics based on motion compensation; and denoising the sub-blocks with UMS characteristics by adopting a spatial domain filtering method based on local spatial domain correlation, and obtaining a spatial-temporal combined denoising result of the lead-zinc flotation foam image to be processed according to the correlation coefficient of the bubble sub-blocks, the combined time domain filtering result and the spatial domain filtering result. Experimental verification is carried out in visual monitoring of the lead-zinc flotation process, the result shows that the foam image with the high signal-to-noise ratio can be obtained through the method, the structural similarity of the denoising result is high, and a foundation is laid for accurate extraction of the visual characteristics of the lead-zinc flotation foam and effective monitoring of the flotation process.

The invention discloses a cuboid surface distortion correction method, which includes: firstly utilizing an image acquisition system to acquire a rectangular image, finding four edges of a cuboid surface through edge detection, intercepting an area defined by the four edges, and then interpolating the intercepted area defined by a trapezoid to obtain a correction image. By means of determination of a vanishing line of the image, the cuboid surface distortion correction method interpolates the image proportionally in the vanishing line direction and the perpendicular direction of the vanishing line to complete correction of projection transformation and affine transformation of the image finally. The method only needs one image and can complete restoration of the original rectangular image. Length-to-width ratio of the corrected image is the same as that of an actual rectangular face.

The invention discloses an image denoising method based on level set curvature and wavelet transform, includes: performing wavelet transform stratification processing for high-density flexible substrate image containing noise, S2 extracting high-frequency component images and low-frequency component images respectively from the sliced images, S3 establishing an image level set curvature variational smoothing model for image smoothing, S4 fusing the high-frequency image after smoothing with the low-frequency image after slicing, and reconstructing to obtain the denoised image. Compared with other denoising models, the novel method of the invention has higher peak signal-to-noise ratio and higher structural similarity.

In order to solve the technical problems that a traditional lens-free camera image reconstruction method is relatively sensitive to noise and relatively low in system depth of field, the invention provides a lens-free camera image reconstruction method based on a coding mask and a Learned-TSVD algorithm. The method comprises the following steps: encoding a propagation process of light by using an encoding mask, converting an original large-scale system measurement matrix into a left system measurement matrix and a right system measurement matrix which are small in scale by utilizing the separable characteristic of the coding mask and a TSVD algorithm; thirdly, constructing neural network training to circularly train the left and right system measurement matrixes, and reducing an error of an approximate operation on a final result; and finally reconstructing an image through the TSVD algorithm and a regularization algorithm. According to the method, the learned system measurement matrixes are used for subsequent calculation, so that the noise influence resistance of the whole reconstruction process is higher; scene images at other distances can be well reconstructed by using the learned system measurement matrixes, and the problem of low depth of field of other reconstruction algorithms is solved.

The invention discloses an improved B-spline non-rigid registration method for distortion correction of an atmospheric turbulence image, relates to the field of atmospheric optics, and mainly solves the problem that the correction performance is unstable when the atmospheric turbulence image is seriously distorted in a traditional registration method. The method comprises the following steps: inputting an image sequence degraded by atmospheric turbulence, taking an average frame of the image sequence as a reference image, and taking each frame in the image as a floating image; generating a uniform grid control point matrix; generating an image deformation field by adopting B-spline non-rigid transformation; the floating image is transformed according to the deformation field, and normalized square difference between the floating image and the reference image is carried out to serve as a target function of image registration similarity; solving an optimal control point matrix when the target function is minimum by using an adjustable exponential damping factor LM algorithm fused with a particle swarm algorithm; and obtaining a registered floating image according to the optimal control point matrix. When the method is used for processing multiple frames of distorted turbulence images, the stability is relatively good, and relatively good image correction quality can be obtained.

The invention discloses an image self-adaptive defogging method. The implementation structure comprises an image sensor module, an image self-adaptive defogging module and a general video transmissioninterface module. According to the invention, an image sensor is used for processing collected data through an image self-adaptive defogging module, and encoding and outputting are carried out according to a general video transmission format. The image self-adaptive defogging module can automatically judge input image data, automatically determine whether a scene needs defogging or not, automatically delineate a sky area and execute defogging operation on a pixel area needing defogging. According to the method, the self-adaptive defogging function can be integrated into the equipment, so thatthe capability of the target equipment for collecting information in a foggy scene is improved, and richer original information is provided for subsequent digital image processing.

The invention relates to the technical field of digital image processing, in particular to a local re-coloring algorithm based on image segmentation, which comprises the following steps of: 1, selecting an image to be processed; 2, marking a rectangular area, and enabling the target area to be contained in the rectangular area; 3, executing an image segmentation algorithm to obtain an initial alpha mask of the target area; 4, performing bilateral filtering processing on the initial alpha mask of the target area, and performing binaryzation to obtain a final alpha mask of the target area; 5, the final alpha mask of the target area is input into the image for superposition, and then the image target area can be segmented; 6, inputting the image target area into an image re-coloring algorithm; 7, obtaining an RGB color convex hull and a corresponding RGBA layer; 8, moving the peak of the RGB color convex hull to change the color of the corresponding image layer, and obtaining a re-colored image target area; and 9, overlapping to obtain a local re-colored image. According to the invention, local re-coloring can be well realized.

The present invention provides a method of lowering intraocular pressure which comprises administering a therapeutically effective amount of a pharmaceutical composition comprising 4-bromo-5-(2-imidazolin-2-ylamino)benzimidazole, or a salt thereof to the affected eye of a patient, as a single dose, wherein the affected eye has an intraocular pressure less than the baseline.