61results about How to "Improve visual quality" patented technology

The invention discloses an image fusion method based on a double-branch feature decoupling auto-encoder, and belongs to the field of computer vision and image processing. According to the method, for the problems of modal pollution and structural distortion in infrared and visible light image fusion, structural semantic information and high-frequency texture details of a source image are extracted respectively by constructing a content feature coding module and a detail feature coding module, and feature decoupling is achieved. And performing deep fusion on the decoupled features by using an adaptive feature weighting mechanism, and reconstructing a fused image with infrared target saliency and visible light detail definition through a shared decoder. According to the method, an end-to-end two-stage training strategy is adopted for optimization, complex prior or post-processing is not needed, the effects of improving the fused image contrast, edge preservation and target recognition performance are remarkable, and the method is suitable for the fields of weak light monitoring, intelligent perception, unmanned system navigation and the like.

The invention provides an image mirror surface highlight removal method and device based on a gradient mask and electronic equipment. The method comprises the following steps: acquiring a mirror surface highlight image and prior semantic information thereof; based on the prior semantic information, performing mirror surface highlight detection on the mirror surface highlight image to obtain a highlight detection result; according to the highlight detection result and the gray scale and illumination information of the mirror surface highlight image, constructing a gradient mask used for representing highlight intensity change; and based on the gradient mask and the prior semantic information, performing image restoration on a highlight area of the mirror surface highlight image, and performing adaptive fusion on a restoration result and a non-highlight area in the mirror surface highlight image to obtain an output image after mirror surface highlight is removed. According to the technical scheme, by introducing the prior semantic information and the gradient mask, accurate detection and smooth restoration of the mirror surface highlight area are realized, and the visual quality of the image can be improved.

This invention provides a method for generating city-level real-scene 3D geographic scene data, relating to the field of 3D modeling and analysis technology. The method includes: collecting multiple sets of historical defect sample records; identifying multiple modeling defect areas; extracting multiple structural deviation features of the modeling defect areas related to different model construction stages to construct a defect structure association model for each modeling defect type; extracting scene stage association data and constructing a scene stage association list; after determining the target defect area of ​​the target scene model, identifying the target defect type and target scene type of the target defect area; constructing scene association tracing paths and stage association tracing paths for the target defect area; fusing the scene association tracing paths and stage association tracing paths to perform source tracing optimization of the target scene model, generating real-scene 3D geographic scene data corresponding to the target scene model. This invention improves the construction quality and efficiency of city-level real-scene 3D data.

This invention relates to the field of image processing technology, specifically to an AI image processing method for electronic rearview mirrors. In this invention, an original rearview mirror image is acquired and decoupled analysis is performed based on a physical degradation model to generate a color attenuation component and a detail interference component. The color attenuation component is processed by a global color and illumination reconstruction network, and a scene semantic attention map is generated using an embedded scene semantic attention module to guide adaptive reconstruction, outputting a first optimized image. Simultaneously, the detail interference component and the scene semantic attention map are input into a local structure to guide a detail repair network, and differential adjustment is performed to output a second optimized image. Pixel-level fusion weights are generated using the scene semantic attention map as a guide, and the two optimized images are weighted and fused to output the final optimized image. This method achieves adaptive and differential intensity control of different semantic regions through the scene semantic attention map, solving the problem of inappropriate regional effects caused by uniform processing of the entire image.

This invention discloses a lightweight zero-reference underwater image enhancement method and system for yellow-green turbid water bodies, belonging to the field of image processing technology. It solves the problems of existing technologies lacking dedicated correction mechanisms for channel imbalances and local turbidity inhomogeneities, which easily lead to red-purple overcompensation and particle noise amplification. The method includes preprocessing the original yellow-green turbid underwater image and performing color-turbidity feature analysis. The pre-corrected image is input into a lightweight zero-reference curve parameter estimation network, and the pre-corrected image is iteratively enhanced multiple times to obtain the underwater enhanced image. By introducing color-turbidity feature analysis and an adaptive correction mechanism, this invention achieves accurate modeling of the unique channel imbalances and local turbidity inhomogeneities in yellow-green turbid water bodies under zero-reference conditions. Simultaneously, relying on the lightweight zero-reference curve parameter estimation network, it significantly improves the visual quality of underwater images in yellow-green turbid environments and the reliability of downstream visual tasks.

The embodiment of the invention provides an out-of-focus lens and a design method thereof. N annular optical correction areas and N annular optical out-of-focus areas are alternately arranged on the radial outer side of one circular optical correction area and are concentric. The ratio of the total area of the annular optical defocus area to the total area of the defocus lens is 0.55-0.58. The focal power of the optical correction area is first focal power, and the focal power of the optical defocus area is second focal power. The optical correction area image space focus is located on the image space of the annular optical correction area image space focus. By applying the embodiment, the plurality of optical correction areas form a clear image on the retina of the human eye to correct refractive error of the human eye, the plurality of optical defocus areas form an image in front of the retina to generate myopia defocus to delay eye axis growth, the optical correction areas and the optical defocus areas are alternately arranged, and the eye axis growth is delayed by controlling the areas of the optical correction areas and the optical defocus areas. The peripheral defocusing effect is achieved, the visual quality is improved, and the myopia progress is delayed.

The application discloses an infrared image enhancement method, system, device and storage medium based on a wavelet threshold, which comprises the following steps: performing multi-scale discrete wavelet transform on an input original infrared image to separate low-frequency components and high-frequency components; the low-frequency components bear overall outlines and illumination distribution information of the image, and the high-frequency components contain edge, texture details and noise information of the image; performing improved adaptive wavelet threshold denoising processing on the high-frequency components to obtain denoised high-frequency components; performing layered enhancement processing based on weighted guided filtering and multi-scale Retinex on the low-frequency components to obtain enhanced low-frequency components; performing wavelet inverse transform reconstruction on the enhanced low-frequency components and the denoised high-frequency components to obtain a preliminary enhanced image; performing size upsampling on the preliminary enhanced image, and applying adaptive contrast restriction histogram equalization processing to output a final enhanced infrared image. The method can effectively suppress noise, retain details, adaptively enhance low-resolution infrared images, and is computationally efficient.

The present application belongs to the field of CT tomographic reconstruction technology and artificial intelligence, and discloses a limited-angle CT reconstruction artifact suppression method based on a multi-domain feature fusion network. In view of the problem that the reconstruction result of traditional CT scanning under limited-angle conditions is prone to artifacts and structural distortion, thereby affecting the image quality and defect detection accuracy, the present application constructs a multi-domain feature fusion artifact suppression network, takes the limited-angle reconstruction result as input, and realizes artifact suppression and detail recovery through the synergistic effect of the encoder part, the decoder part, the feature enhancement part and the feature conversion part. The present application can obtain high-quality tomographic images under limited-angle conditions, effectively reduces the scanning angle and time of industrial CT detection, improves the imaging clarity and reliability without increasing the radiation dose, is suitable for industrial detection of complex structure workpieces, and has important industrial application value.

The application discloses a model training method, a poster image processing method, an electronic device and a storage medium. The model training method comprises the following steps: obtaining training sample data, wherein the training sample data comprises first sample data and second sample data, the first sample data comprises an obtained actual poster image and a text mask of the poster image, and the second sample data comprises a poster image obtained by generating poster text by using an image, a text mask of the poster image and an original image for generating the poster image; and training a first model by using the training sample data, wherein the first model is used for removing text in a poster image to be erased and repairing a text area.

The embodiment of the application discloses a low-illumination image enhancement method and device based on an improved normalized flow, which comprises the following steps: acquiring a low-illumination image to be enhanced; performing normalization processing on the low-illumination image to be enhanced to obtain a normalized spectrum; performing edge extraction processing on the normalized spectrum to obtain an edge spectrum; performing feature fusion on the low-illumination image to be enhanced, the normalized spectrum and the edge spectrum to obtain a feature fusion image; performing feature extraction on the feature fusion image by using a feature extraction module to obtain a feature extraction image; performing feature mapping on the feature extraction image by using a feature mapping module to obtain a feature mapping image; and inputting the feature mapping image into a normalized flow reversible network to output a normal-illumination image. The low-illumination image is enhanced by the method, and the enhancement effect can be improved.

This invention discloses a specular removal method based on grouped enhanced convolutional attention feature fusion, belonging to the field of image processing technology. The method constructs a specular removal model based on a CycleGAN network, which consists of a generator, a discriminator, and a loss function optimization module. The generator includes detail enhancement depth downsampling, upsampling, and grouped enhanced convolutional attention fusion modules, effectively capturing image details, reducing checkerboard artifacts, and enhancing feature fusion and capture capabilities. The discriminator uses a PatchGAN structure to improve local detail capture capabilities. The optimized loss function integrates multiple losses to improve model performance. Experimental results show that compared with various traditional and deep learning methods, the method of this invention outperforms in PSNR and SSIM metrics, exhibits strong adaptability, and can provide high-quality image data for subsequent detection of elevator door components.

The invention discloses a complex-material-oriented multi-path full-frequency global illumination neural rendering method and system, and belongs to the technical field of computer graphics, and the method comprises the steps: carrying out the scene query of light emitted by pixels, and obtaining a primary geometric buffer and a multi-path mixed geometric buffer of the multi-time ejection of the light; respectively extracting main frequency features and multi-path sub-frequency features in a neural feature field based on the two features; performing directed graph structured modeling based on the multi-path sub-frequency features to obtain propagation features representing light path propagation information and global features representing overall light path information; performing multi-path frequency joint feature fusion based on the dominant frequency feature, the propagation feature and the global feature to generate a fusion feature with a full-frequency feature; and estimating a final radiation brightness value of the pixel based on the fusion feature, and completing neural rendering of global illumination. According to the method, efficient modeling and high-quality drawing of multi-path full-frequency global illumination under a complex material can be realized, and the sense of reality and dynamic consistency of a rendered image are remarkably improved.

The application relates to the technical field of underwater robot vision processing, and discloses a GAN-based underwater robot image real-time repairing and target identification method. The method acquires an original image sequence obtained by an optical sensor of an underwater robot, extracts degradation features of each frame of image, including scattering noise distribution, color distortion parameters and motion blur intensity. The original image sequence is input into a network, and based on the degradation features, a repaired image is generated frame by frame, and local structure consistency features and global semantic coherence features of the repaired image are extracted. A cascaded classifier is used for multi-scale target detection of the repaired image, local structure features and global semantic features are combined, and a target boundary box and a category label are output. According to the spatial distribution density of the target boundary box, the priority of the repaired area is dynamically adjusted, and the synchronous output of the repaired result and the target identification result is realized. The method can effectively improve underwater image quality and target identification precision.

The invention discloses a remote sensing image robust watermarking method based on an improved reversible neural network, and relates to the technical field of digital image copyright protection, and the method comprises the steps: collecting an original remote sensing image and a binary watermark sequence, and carrying out the preprocessing, and forming a training sample; inputting the training sample into the watermark model; applying randomly selected noise attack to the watermark image through a noise layer to generate a disturbed watermark image; inputting the disturbed watermark image into the watermark model, and executing a reverse extraction path to obtain extracted watermark information; calculating a multi-target loss function, wherein the multi-target loss function is based on the original remote sensing image, the binary watermark sequence, the watermark image and the extracted watermark information; and training the model until convergence. Through a robust lifting mechanism and an improved reversible neural network structure, the method shows extremely high stability to various complex noise attacks, and has good generalization ability and environmental adaptability.

The invention provides an RGB-D image salient target detection method based on frequency decoupling mode interaction, and the method comprises the steps: firstly carrying out the multi-stage feature extraction of RGB-D mode data, and obtaining the multi-stage feature representation of an RGB-D mode; performing frequency domain sensing cross-modal interaction on the RGB-D modal multi-stage feature representation to obtain frequency sensing cross-modal spatial domain interaction features; discriminative enhancement and cross-modal fusion are carried out on the cross-modal spatial domain interaction features of frequency sensing, and final multi-modal fusion features are obtained; and based on the final multi-modal fusion features, through multi-scale aggregation and global dependence modeling, generating a saliency target detection prediction map. According to the method, cross-modal association in a frequency domain is explicitly modeled in a feature learning process by using a frequency cross Mama fusion module, so that the problem that internal relationships among modals may be ignored when fusion is directly performed in a spatial domain in a traditional method is remarkably relieved, and the performance and generalization ability of salient target detection are improved.

This utility model discloses a novel SMT placement equipment, comprising: a placement equipment body, through which the working status of the placement equipment body can be observed via an arc-shaped observation window on a protective cover; a motor on the protective cover rotates a lead screw to move an LED light within the protective cover, providing more precise illumination and ensuring that the area observed within the observation window is illuminated; simultaneously, the extension and retraction of a pneumatic telescopic rod allows the LED light to be adjusted in angle, achieving optimal lighting effects and providing a better user experience; when the inner side of the arc-shaped observation window fogs up or becomes dusty, the pneumatic telescopic rod can be extended and retracted to bring a cleaning wipe into contact with the inner side of the arc-shaped observation window; the cleaning wipe is then moved along with the movement of a fixed frame to wipe the inner side of the arc-shaped observation window, providing a clearer and higher visual quality; after wiping, the cleaning wipe is retracted to prevent it from continuously rubbing against the observation window while the fixed frame is moving.

The invention discloses a low-illumination image dynamic noise reduction method for a daytime effect night vision device based on image recognition, and the method comprises the following steps: obtaining a low-illumination image sequence, and determining a current frame as a target frame; performing noise level estimation and image recognition on the target frame to obtain a noise parameter and a recognition confidence map; generating an initial edge set based on the target frame, and determining an initial edge weight; performing semantic boundary constraint reconstruction on the initial edge set, and outputting a reconstructed edge set; performing Felzenszwalb graph segmentation on the reconstructed edge set, sorting the reconstructed edge set according to the reconstructed edge weight, taking edges in sequence according to the sorting sequence, and outputting a final region label graph and a segmented region set according to a merging criterion; determining a region-level noise reduction intensity parameter for each region, and generating a noise reduction control chart; and performing regional dynamic noise reduction processing on the target frame according to the noise reduction control chart to obtain a noise reduction output frame. According to the method, the target edge and detail fidelity is improved, and the excessive smoothing risk of cross-semantic boundaries is reduced.

The utility model relates to the technical field of intelligent glasses, and discloses a myopia prevention and control lens and intelligent glasses, which comprise a glasses frame, two ends of the glasses frame are rotatably connected with a glasses frame, the outer wall of the glasses frame is fixedly connected with an antiskid sleeve, one side of the outer wall of the glasses frame is fixedly connected with a lens, and the middle part of the outer wall of the glasses frame is fixedly connected with a nose bridge frame. A protection assembly is arranged on the outer wall of the lens; the protection assembly comprises a point diffusion frosted layer and a hollow part, the point diffusion frosted layer is arranged on one side of the outer wall of the lens, and the hollow part is arranged in the middle of the point diffusion frosted layer. According to the utility model, the development of myopia is effectively prevented through the point diffusion frosted layer and the gradient change defocus design, meanwhile, the outer wall of the lens is covered with the double-sided antireflection film and the high-transmittance and high-cleanliness UV film, the light transmittance is improved, ultraviolet rays are blocked, eyes are protected from being damaged, and the lens adopts the design of thin middle part and thick edge, so that the visual quality is improved, and the comfort level is enhanced; therefore, myopia prevention and control are realized.

The application discloses a kind of corrective lens design equipment, storage medium and lens, it is related to medical equipment field, for solving the problem of existing refractive correction improves visual acuity or the effect of improving visual quality is poor.Refractive lens design equipment includes acquisition device, selection device and design device, acquisition device obtains the central refractive power of target individual eyeball, selection device finds out target refractive topography from refractive topography group, central refractive power is in the central refractive power range corresponding to target refractive topography, and refractive topography describes the retinal refractive power distribution of eyeball with corresponding central refractive power in preset field angle;Design device obtains refractive power distribution correction scheme in conjunction with target refractive topography, to obtain the corrective lens for target individual.The present application combines the peripheral defocus characteristics of human eye retina to design semi-personalized lens for individual, to achieve the purpose of improving visual acuity, improving myopia prevention and control effect or improving peripheral visual quality.

The invention discloses a personalized image generation method and system based on reusable model library retrieval. An image generation instruction is encoded through multi-modal embedding generation to obtain a multi-modal query embedding vector. And performing screening in a pre-constructed model library by using the multi-modal query embedding vector to obtain at least one target reference check point most related to the image generation instruction and a candidate adapter, and performing semantic correlation reordering secondary screening on the candidate adapter to obtain a target adapter. And constructing model configuration for image generation according to the screened target reference check point and the target adapter, and generating or editing the input noise or the reference image based on the model configuration to obtain a personalized image. According to the method, diversified and long-tail style generation requirements can be covered without additional training, the personalized instruction execution capability is effectively improved, the performance of an existing large model on style blind spots, role consistency and other problems is improved, and the method has the advantages of being high in adaptability, high in generation quality and good in expansion efficiency.

This invention discloses an adaptive dual-image reversible information hiding method based on a Sudoku matrix. First, an original carrier image is copied twice, and a Sudoku reference matrix is ​​established. Next, the secret information of the binary random bit string is divided into groups of 6 bits each, and each group is converted into two octal secret numbers. A pair of pixel values ​​from the original carrier image is mapped onto the Sudoku reference matrix, establishing a quantization distortion model. Then, based on the quantization distortion model, the optimal distance threshold and optimal embedding strategy are determined, and two cryptic images are generated. Finally, two receiving ends establish the same Sudoku reference matrix as the sending end, recover the pixel pairs of the original carrier image, traverse all pixel pairs of the cryptic images, and convert the extracted secret numbers into binary bit strings. This invention significantly improves image visual quality, is very simple and practical to operate, and effectively enhances the security of information transmission.

This disclosure provides a method and system for enhancing low-exposure images based on embedded McLaurin and convolution, relating to the field of image processing technology. The method involves acquiring a low-exposure image to be enhanced, processing and decomposing the low-exposure image, inputting the low-exposure image into a backbone network to extract multi-scale features, introducing a dynamic adjustment factor into the backbone network, using the dynamic adjustment factor to learn a convolution weight calibration map, calibrating the convolution weights of the backbone network with the learned weight calibration map, performing convolution operations on the features using the calibrated convolution kernels to obtain the luminance component of the low-exposure image, and outputting a long-exposure enhanced image that varies with the dynamic adjustment factor using the luminance component.

The invention discloses a double-stage document image general restoration method and system for a large model, and the method employs a double-stage restoration model for image restoration, and comprises a space correction stage and a pixel correction stage. In the spatial correction stage, the degraded image is processed through a double-branch structure of a layout branch and a text line branch; wherein the layout branch predicts a sparse two-dimensional mapping field and a three-dimensional coordinate grid to recover a global geometric structure, and the text line branch generates a local modulation signal based on text region features extracted by a pre-training text segmentation model; fusing global and local information, generating a dense two-dimensional mapping field, and outputting a space correction image by using reverse sampling operation; in a pixel correction stage, a spatial correction image and a gradient map thereof are used as conditions to be input into a denoising diffusion implicit model architecture to gradually recover pixel values. According to the method, general repair of various degradation problems is completed through double-stage repair, so that the content reading capability of various large models on document images is improved.

The application discloses an infrared image non-uniformity prediction and correction method based on wavelet transform, which comprises the following steps: adopting double-density dual-tree complex wavelet transform to perform multi-scale decomposition on an input infrared image; in a high-frequency subband, detecting a blind element position based on a local variance statistical method; in a low-frequency subband, constructing an autoregressive model, predicting non-uniformity of a background region, and performing accurate processing on a prediction result; combining the prediction results of the high-frequency subband and the low-frequency subband to generate a pre-correction coefficient; and adopting the pre-correction coefficient to perform correction processing on an original infrared image to obtain a corrected image. Through the combination of an adaptive fusion strategy, space-time consistency constraint and dynamic gain and bias updating technology, the problems of non-uniformity noise, balance between details and background, device drift compensation and consistency in a dynamic scene in infrared image processing are solved, and the accuracy and stability of infrared image processing are significantly improved.

The present application relates to the field of image processing and computer vision, and discloses an unsupervised low-light image enhancement method and system, the method comprising the following steps: S1, constructing a diffusion model; S2, training the diffusion model using pairs of low-quality images in a public dataset and the sum of four loss functions; S3, training the diffusion model using any non-paired low-light image, normal-light image and the sum of two loss functions; S4, using the reflectance map of the low-light feature and the illumination map of the normal-light feature as the diffusion model, and under the guidance of the low-light feature, the enhanced feature is obtained by restoration, and the enhanced feature is used as the input of the decoder to reconstruct the final enhanced image. The system comprises a model construction unit, a model training unit and an enhanced image output unit. The present application also discloses an electronic device and a computer readable storage medium. The present application is used for low-light image enhancement by using traditional retinal theory and deep neural network.

The invention relates to the technical field of image processing, and discloses an operation scene real-time dynamic reconstruction method and system based on three-dimensional Gaussian sputtering. The method comprises the following steps: extracting spatial-temporal characteristics of a soft tissue deformation Gaussian subset through a HexPlane spatial-temporal encoder; fusing the decoded feature vectors to generate a soft tissue deformation parameter set; a deformation field Jacobian matrix of Gaussian elements in the soft tissue deformation Gaussian subset is calculated, a local volume strain rate is obtained through determinant operation, when the strain rate exceeds a preset threshold value, self-adaptive densification operation is executed, and new Gaussian elements inherit parent Gaussian element deformation parameters and are updated to a corresponding set; and combining the updated parameters to carry out probabilistic boundary allocation and transmissivity perception rendering to obtain a reconstruction result. The accuracy, the physical authenticity and the stability of real-time dynamic reconstruction of the operation scene are improved.

This disclosure provides a virtual try-on method, apparatus, device, and storage medium. The method includes acquiring an original image; extracting an original clothing region image from the original image using a target segmentation model; performing outward cropping processing on the original clothing region image to obtain a cropped clothing image and adjusting its resolution to obtain a clothing image to be processed; determining a first latent space feature matrix corresponding to the clothing image to be tried on; merging the first latent space feature matrix with a second latent space feature matrix determined from the target clothing image to be processed; denoising and redrawing the merged latent space matrix to obtain a new merged latent space matrix; determining a target clothing region image from the target second latent space feature matrix; and replacing the original clothing region image in the original image with the target clothing region image to obtain a target try-on image. This disclosure only processes the image corresponding to the clothing region, preserving the clarity and detail information of the original image to the greatest extent.

The image motion deblurring method based on a hybrid expert network belongs to the technical field of image processing and computer vision, and particularly relates to an image motion deblurring method based on a hybrid expert network; solves the technical problems of coarse-grained routing mechanism, sparse activation detail loss, imbalance between lightweight and precision, single feature fusion strategy, and difficulty in balancing precision and efficiency in the prior art; the method comprises the following steps: inputting a single-frame blurred image to be repaired into a hybrid expert deblurring model that has been trained, and reconstructing and outputting a clear image after deblurring. The image motion deblurring method based on a hybrid expert network is suitable for the fields of lightweight image visual perception, image restoration and deblurring processing.

The invention discloses an underwater image enhancement method and system based on physical prior and target features, and belongs to the technical field of image processing. In order to solve the technical problem that an existing underwater image enhancement method is difficult to consider imaging physical consistency, visual quality and downstream target identification task availability at the same time, physical prior information is extracted based on an underwater image, and the physical prior information comprises a transmissivity image and a background light image; obtaining target features based on the underwater image; extracting image features of the underwater image through an enhancement network, and fusing and reconstructing the image features, the physical prior information and the target features to obtain an enhanced image; and training the enhancement network, and processing the underwater image by using the trained enhancement network to obtain an enhanced image. According to the method, the purposes of realizing high-quality visual restoration on the premise of meeting underwater imaging physical constraints and remarkably improving underwater target recognition precision and related downstream task performance can be achieved.