104results about How to "Enhance image detail" patented technology

The invention provides a face detecting and tracking method and a device. The method comprises the steps of inputting a face image or a face video, preprocessing the face image or the face video in an illumination mode, detecting a face by usage of an Ada Boost algorithm, confirming an initial position of the face, and tracking the face by the usage of a Mean Shift algorithm. According to the face detecting and tracking method and the device, a self-adaptation local contrast enhancement method is provided to enhance image detail information in the period of image preprocessing, in order to increase robustness under different illumination conditions, face front samples under different illumination are added to training samples and accuracy of the face detection is increased by adoption of the Ada Boost algorithm in the period of face detection, in order to overcome the defect that using color of the Mean Shift algorithm is single, grads features and local binary pattern length between perpendiculars (LBP) vein features are integrated by adoption of the Mean Shift tracking algorithm in the period of face tracking, wherein the LBP vein features further considers using LBP local variance for expressing change of image contrast information, and accuracy of the face detection and the face tracking is improved.

The invention relates to an underwater image enhancement method based on dark channel prior algorithm and white balance. The method achieves better image enhancement through the modification of background light, improves the image definition and contrast, and achieves the purpose of image optimization. The employed technical scheme is that the method comprises the following steps: A, background light processing step; B, dark channel prior step: 1, obtaining a fined transmissivity t(x) after image cut and a defogged image through employing the dark channel prior algorithm; 2, exporting a depth map d(x) through the transmissivity t(x); 3, obtaining a needed mask; 4, extracting an AOI (area of interest); C, white balance step. The method is mainly used in an occasion of underwater image enhancement.

A process and methods of digital enhancement of motion pictures and other moving image sequences for the purpose of being exhibited in an alternative display format including a large format cinema are disclosed. The invention efficiently enhances image resolution and quality through a temporal filtering process and achieves high performance using automated or interactive statistical quality evaluation methods. A system specially designed for efficient temporal computing with a parallel and distributed computing configuration equipped with a variety of optimization schemes is also disclosed. The performance of the process and the system is optimized through an intelligent controller and is scalable to support any throughput requirements demanded for concurrent motion picture releases in the original format as well as in any alternative format.

The method comprises: shooting multi pictures with different exposure levels at same scene; according to the response curve of camera, portioning the each picture into several blocks; selecting the blocks having the most information to make block combination; using a fusion function to remove the edge effects between blocks to finally get an enhanced picture with wide dynamical range.

The invention discloses an image defogging transmittivity optimization method based on dark channel prior. The method comprises the following steps that firstly, rough transmissivity is solved according to the dark channel prior theory, a foggy image is subjected to rough defogging, and an intermediate defogging result is obtained; secondly, the transmissivity is optimized through the intermediate defogging result so that the transmissivity can be more approximate to a true transmissivity value, and then steering filtering is used for refining the optimized transmissivity to eliminate the blocking effect. According to the method, the transmissivity is optimized, the optimized transmissivity is more approximate to the true value of a scene, and the color cast phenomenon appearing in some special areas through a traditional dark channel prior algorithm is relieved. The improving method aims at optimizing the transmissivity and is quite simple and effective, details of the restored image are greatly enhanced, the overall brightness is improved, the visual effect is better, and the defogged image is more lifelike.

The disclosure describes a technique for medical imaging, referred to herein as the Rapid Interleave Overlap Technique (RIOT), wherein image data is acquired as a plurality of series sequences in a manner that allows for unlimited overlap. RIOT involves interleaving and overlapping 2D image slices of multiple series of image data of the same ROI into a composite data set from which MPR and 3D reconstructions exhibiting excellent resolution properties and crisp image quality can be generated.

The invention provides a rapid image splicing method based on feature matching. The method includes the steps: respectively extracting line features and point features of images by the aid of a Canny edge detection algorithm and a Harris corner point detection algorithm and combining the line features and the point features to obtain the best feature points; roughly matching the feature points by the aid of similarity metric NCC (normalized cross correlation), removing mismatched points by the aid of an RANSAC (random sample consensus) algorithm to improve image matching accuracy, and calculating transformation model parameters by an LSM (least square method); finally, fusing the spliced images by a weighted average method and eliminating splicing gaps. The points to the matched are determined according to the point and line image features, image details can be enhanced, image matching errors caused by underexposure, overexposure, camera shake and the like are avoided, and image splicing quality is improved to a certain degree.

The invention discloses an image defogging method facing to an atmospheric scattering proximity effect. The method comprises the following steps: 1) obtaining the image transmittivity of a primary image and obtaining an initial defogging image of the primary image through an image defogging algorithm; 2) establishing a radiation transmission equation of the image according to the image transmittivity of the primary image; 3) solving the radiation transmission equation to obtain an atmospheric point diffusion function; 4) performing the deconvolution processing on the initial defogging image according to the atmospheric point diffusion function to obtain a final defogging image. The method provided by the invention can be used for reducing the influence of the atmospheric scattering proximity effect on the imaging image, improving the contrast and definition of the image, increasing the image detail and achieving a good image defogging effect.

The invention discloses an image super-resolution reconstruction method based on a multi-band deep convolutional neural network, comprising the following steps: selecting training samples and test samples; inputting training images to the network in batches for feature extraction, and combining long-term and short-term memories to construct a multi-band learning structure, then performing featurerefinement, feature mapping and up-sampling reconstruction; and obtaining network parameters according to the model obtained through training to complete the image reconstruction. The invention enhances high-frequency information reconstruction by introducing multi-band learning, and uses the memory migration operation to make the network have long-term and short-term memories at the same time, which speeds up the image reconstruction speed, enhances image edge and texture detail reconstruction, and obtains image super-resolution reconstruction results of better quality. The invention has strong super-resolution capability, and the reconstructed image is closer to the real image.

The invention discloses a dark channel prior and variation Retinex-based underwater image enhancement method. The underwater image enhancement method comprises the following steps of based on the meanvalue and the variance of each channel, linearly stretching on an original degraded underwater image in the RGB color space so as to correct the color cast; subjecting the image after being subjectedto color cast correction to defogging treatment based on the dark channel prior theory, removing the influence of backward scattering, and improving the contrast ratio of the image; converting the processed image from the RGB color space to the lab color space, and solving the uneven illumination problem on an L component by using a variable Retinex model; recombining the enhanced L component with a component a and a component b, and converting the image back to the RGB color space to obtain a final enhanced image. According to the invention, the problems of color degradation, low contrast ratio, fuzzy details and uneven illumination in an underwater image are solved. Meanwhile, the visual quality of underwater degraded images is improved.

A prepress workflow process suitable for generating a reproduction composite image from a source image that may be tolerant to or exhibit improved tolerance to misregistration when printed is provided. Embodiments of the prepress workflow process include obtaining a source image. Next, N≧2 color separations are generated from the source image. The color separations may be process or non-process color separations. The N color separations are first processed globally for improving the misregistration tolerance of a reproduction composite image to be formed therefrom. If the misregistration tolerance of the reproduction composite image is not acceptable to be user, at least a portion of each N color separation may be further processed at a local level for producing an improved misregistration tolerant reproduction composite image.

The invention provides a remote sensing image defogging method under the inhomogeneous cloud and fog condition. The remote sensing image defogging method includes the following implementation steps: (1) inputting a foggy remote sensing image, conducting inhomogeneous cloud and fog preprocessing, and obtaining a degraded image evenly influenced by fog; (2) obtaining the transmittance and the daylight value of the evenly-degraded image through a basic defogging algorithm; (3) solving an atmospheric point spread function of the image through the adoption of generalized Gaussian distribution according to the transmittance; (4) obtaining a final defogged image in cooperation with an atmosphere multiple scattering image degradation model through the adoption of the obtained daylight value, the obtained transmittance of the degraded image and the atmospheric point spread function of the image. As for the problem of remote sensing image degradation caused by inhomogeneous cloud and fog, inhomogeneous cloud and fog removing is achieved through the adoption of image information, the image transmittance is obtained from the image, the atmospheric point spread function is solved, and finally image final defogging is achieved through the atmosphere multiple scattering image degradation model; the contrast ratio and the definition of the image are improved, and details of the image are increased.

The invention discloses an infrared image enhancement method and device based on texture weighted histogram equalization, equipment and a computer readable storage medium. The method comprises the steps of determining a local extremum difference of each pixel point in an original infrared image; comparing the local extremum difference of each pixel point with a preset difference threshold, and recording the positions of the pixel points of which the local extremum differences are greater than or equal to the preset difference threshold in the original infrared image to obtain a statistical area; performing histogram statistics on the original infrared image in the statistical area to obtain a statistical area histogram; carrying out nonlinear transformation on the histogram of the statistical region to obtain a histogram after nonlinear transformation; performing cumulative probability distribution calculation on the histogram after nonlinear transformation to obtain a gray mapping function; and inputting the original infrared image into the grayscale mapping function, and outputting a target enhanced infrared image. According to the method, the device, the equipment and the computer readable storage medium provided by the invention, the phenomena of background overenhancement and local noise amplification can be inhibited.

The invention provides a homomorphic filtering based image processing method and system. The method includes the steps of acquiring an original code map; subjecting the original code map to homomorphic filtering to obtain a code map subjected to homomorphic filtering; based on a preset partitioning algorithm, partitioning the code map subjected to homomorphic filtering into a plurality of code map sub-units; subjecting each code map sub-unit to binarization by an OTSU algorithm to obtain a binary code map sub-unit; combining the binary code map sub-units to generate a final binary code map. The homomorphic filtering based image processing method and system has the advantages that before binarization of the code map, homomorphic filtering is performed first, the images are effectively enhanced by homomorphic filtering, image quality decline caused by insufficiency of light is effectively avoided, the advantages of the OTSU algorithm are given to maximum play, a foreground and a background are precisely distinguished, no effective information is lost from the binary images, image binarization quality is improved, and the following image processing is simplified.

The invention discloses an infrared image enhancement method. A histogram equalization method is used to process an image; a maximal inter-class variance method is used to obtain a segmentation threshold for segmenting the image into background and objects, an image a background area, and an image area greater than or equivalent to the segmentation threshold serves as an object area; effective grayscale levels in the background and object areas are counted and ordered; and according to the order of the grayscale levels, the grayscale levels of the background and object areas are compressed and extended within certain grayscale level range in an equidistant way, and an enhanced image is obtained. According to the invention, the background of the infrared grayscale image is compressed, the objects are enhanced, details are emphasized, and the method is suitable for conditions that the object temperature is different from the background temperature.

The invention provides a non-uniform illuminance image enhancement method, and belongs to the technical field of digital image processing. The method aims to solve the problems of an image enhancement method in the prior art; according to the method, image enhancement processing can be carried out on a non-uniform illuminance image. The non-uniform illuminance image enhancement method includes the following steps that a histogram of the non-uniform illuminance image is made, a threshold value is determined according to peak values of high illuminance and low illuminance, the histogram is divided into a high illuminance sub-band and a low illuminance sub-band, a sub-band brightness enhancement method is maintained for the high illuminance sub-band and a corresponding gray level mapping enhancement method is adopted for the low illuminance sub-band. According to the non-uniform illuminance image enhancement method, the situation that the image is excessively enhanced locally is avoided, overall noise of the image is inhibited, and the image which is uniform on the whole is obtained. The calculation amount of an algorithm is little, and the method is stable and reliable and can be used in real-time project systems.

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

The invention discloses an image enhancement method based on non-linear guiding filtering, and belongs to the technical field of image processing. Through establishing a local non-linear guiding filtering model, a corresponding guiding filtering coefficient is solved. In order to further improve efficiency and reduce time complexity, non-linear guiding filtering is performed after subsampling of an image, and then the size of the image is recovered to size of an original image through a bilinear interpolation method. The method can be applied in various different computer vision fields, such as image smoothing, image detail enhancement, HDR image compression, and fog elimination and balancing and other operation, and meanwhile an algorithm is simple and efficient in calculation, and good in realization.

A neural network model training apparatus for enhancing image detail is provided. The apparatus includes a memory and at least one processor configured to obtain a low quality input image patch and a high quality input image patch, obtain a low quality output image patch by inputting the low quality input image patch to a first neural network model, obtain a high quality output image patch by inputting the high quality input image patch to a second neural network model, and train the first neural network model based on a loss function set to reduce a difference between the low quality output image patch and the high quality input image patch, and a difference between the high quality output image patch and the high quality input image patch. The second neural network model is identical to the first neural network model.

The invention provides an enhancing method and system for an infrared image being converted to a pseudo color image in a self-adaptive mode. The method comprises the steps that de-noising processing is conducted on the input infrared image according to median filtering; edge extraction is conducted on the infrared image according to Sobel edge detection; normalization processing is conducted on generated data with enhanced image details, and the data are converted to gray data; histogram statistic and mapping color code article bound calculation are conducted on the gray data; according to a histogram statistic result and color code bounds, color space conversion is conducted on the basis of a mapping rule. Compared with the prior art, the enhancing method and system for the infrared image being converted to the pseudo color image in the self-adaptive mode have the advantages that the purposes of enhancing the image details and reducing image noise are achieved, the dynamic range of the infrared image can be stretched in real time, an observed object is protruded at the same time when the background is compressed, image definition and the image contrast ratio are improved, and the high-quality pseudo color image is output. The image quality is ensured and the imaging effect is greatly improved when the dynamic range of the image is stretched and the image details are enhanced.

The invention discloses a dynamic contrast ratio enhancement method based on function curve transformation, and relates to the technical field of image processing. The method comprises the following steps that a frame image in video is obtained; a gray-scale value of each pixel point in the frame image is obtained; global histogram statistics of the current frame image is established, a gray-scale value a1 is selected, a ratio n is set, the relation between the a1 and the n meets the condition that in the global histogram statistics, the ratio of the sum of the pixel points of which the gray-scale values range from 0 to the a1 to the total pixel points is n, or local histogram statistics of the current frame image is established, and a gray scale average a2 is calculated; the obtained gray-scale value of each pixel point is substituted into a formula, and a corrected gray-scale value of the pixel point is obtained; the gray-scale values of all the pixel points are modified to be I, and an enhanced frame image is obtained; a normalized ratio K is set, a ratio k of a gradient difference before and after adjustment of the current frame image is calculated, if abs(k-K) is smaller than or equal to m, processing of the frame image is completed, otherwise, coeff is adjusted, and the steps from third to fifth are repeated or the coeff is used in a formula of a next frame image.

The invention provides a panoramic image shooting method and a mobile terminal. The method comprises the following steps: respectively acquiring three kinds of images with different exposure degrees so as to synthesize three initial panoramic images; synthesizing a target panoramic image by using three initial panoramic images after high-dynamic range image processing, thereby reducing the occurrence probability of dark and bright stripes on the target panoramic image, improving brightness uniformity of the target panoramic image, and improving the image details and image quality of the targetpanoramic image.

A method and system for defect image correction is disclosed. The present invention comprises a method and system for defect image correction. The present invention generally relates to detecting and correcting defects in one or more digital images caused by occlusions when the image was taken.

The invention discloses a medical image fusion method based on WEMD (Window Empirical Mode Decomposition) and PCNN (Pulse Coupled Neural Network). The medical image fusion method based on WEMD and PCNN comprises the steps: performing a layer of WEMD for a source image, thus not only overcoming the problem that a traditional wavelet method is difficult to select a wavelet primary function and is poor in adaptation, but also effectively avoiding the gray scale speckle phenomenon of the BIMF (Bidimensional Intrinsic Mode Function) component obtained through a traditional BEMD (Bidimensional Empirical Mode Decomposition) method, and accelerating the decomposition speed; designing different fusion rules according to the image characteristics for the corresponding BIMF component and the residual component respectively, and providing visual effect for the fusion image maximumly, wherein the BIMF component utilizes the fusion rule based on an improved and simplified PCNN model to select the clear area of the image and the residual component utilizes the fusion rule based on the area energy to enhance the image detail; and finally obtaining the fusion result by performing WEMD inverse transformation for the fusion component. The medical image fusion method based on WEMD and PCNN solves the problem that in the prior art, the multi-scale decomposition algorithm is poor in adaptation and the edge and the contrast of the obtained image are serious in distortion.

The invention discloses a remote sensing image marine ship identification system and method based on an improved YOLOv4 algorithm. The method comprises the steps of collecting satellite remote sensing images of a sea surface scene shot or collected in the past; performing type labeling on the preprocessed picture by using data labeling software; segmenting a ship in the remote sensing image from a surrounding environment phase to eliminate image noise; obtaining an estimated value of an anchor box of the YOLO algorithm; generating a YOLOv4 framework; generating a detection frame of the YOLOv4; setting a threshold value of the candidate box, and finally obtaining a prediction box; calculating three loss functions and minimizing the total value of the three loss functions to obtain an improved YOLOv4 neural network after training; and inputting the pictures in the test subsets into the trained and improved YOLOv4 network to obtain the target category, the specific position of the target in the pictures and the width and height of the target so as to complete target detection. The sea surface ship target can be rapidly detected and automatically identified, and the ship identification probability and accuracy are high.

The embodiment of the invention provides an image processing method and a device, and a readable storage medium, and the method comprises the steps: firstly determining a first pixel region accordingto a first pixel point in a first image; then, according to the first pixel area and the alignment vector between the first image and the second image under the sub-pixel precision, obtaining N secondpixel areas, corresponding to the first image, in the N second images; fusing the first pixel points and the second pixel points in the second pixel area based on the weight of the pixel distance andthe weight of the pixel value, completing the fusion of the first image and the N second images by traversing each pixel point in the first image, and obtaining a fused image. According to the method, under the sub-pixel precision, the first image and the second image are fused, and the images are filtered from the two aspects of the pixel distance and the pixel value, so that edge details in thefused images are better protected, image blurring is reduced, and a better visual effect is achieved.

The embodiment of the invention discloses an image enhancement method and device, electronic equipment and a medium. The method comprises the steps of determining an upper platform threshold value anda lower platform threshold value of double-platform histogram equalization according to a mean value and a standard deviation of histogram values corresponding to gray values of a base layer image, and performing double-platform histogram equalization processing on the base layer image to obtain a target base layer image; performing gain processing on the detail layer image based on the gain coefficient, and performing piecewise mapping processing based on a piecewise function to obtain a target detail layer image; fusing the target base layer image and the target detail layer image to obtainan enhanced target image, wherein the base layer image and the detail layer image are obtained by filtering the target image. According to the scheme, the problem of excessive stretching of the background can be effectively solved, and details of the image are enhanced while excessive enhancement of the noise of the image is avoided.

The invention discloses a super-resolution image reconstruction method and system based on multi-feature learning, and the method makes full use of the rich information contained in a single input image for reconstruction, and does not depend on an external database. According to the method, a mapping relation between image features is established based on cross-scale similarity of images, and a high-resolution image containing high-frequency information is reconstructed for an input image directly by using the mapping relation, so that the defect of high-frequency information loss caused by image reconstruction by using an interpolation amplification method is well overcome. According to the method, effective high-frequency information is acquired by using singular value thresholding, andthe high-frequency information is amplified by using a gradient feature mapping relation and then is overlapped on a high-resolution image in a blocking manner, so that a final image reconstruction result is obtained. According to the method for reconstructing the image by utilizing the image feature combination, noise points of the reconstructed image are effectively inhibited, image edge and texture information is well kept, and detail enhancement of the image is realized.

The invention provides a Triplet network learning method based on a semantic hierarchical structure, and the method comprises the steps: constructing the semantic hierarchical structure, carrying outthe hierarchical Triplet sampling, carrying out the hierarchical Triplet network training, and carrying out the enhancement through employing a bilinear feature, thereby updating the parameters of a network. According to the method, the semantic knowledge is used for guiding the hierarchical distinguishing of the sample structure of the network, and the relation between the layers is used for enabling the network to pay more attention to more effective Triplets pairs, so that the effectiveness of samples in the batch is fully mined, and the separability of deep features learnt by the network is improved. Meanwhile, image details are enhanced by utilizing a bilinear function, and the image details and Triplet are jointly trained, so that the performance of the network is further improved.