61 results about "Image colorization" patented technology

The invention discloses an interactive grayscale image colorizing method based on local linear model optimization. The method includes: inputting a grayscale image to be processed, converting the grayscale image to be processed into a red, green and blue (RGB) color space input image, conducting a little manual line coloring to obtain a colorized image, converting the grayscale image to be processed and the colorized image to a YUV color space from an original RGB color space respectively, calculating a laplace sectional drawing matrix, optimizing and solving a sparse matrix equation by using a local linear model, obtaining the colorized image based on the YUV color space, and finally converting the colorized image based on the YUV color space into the RGB color space to obtain a final colorized image. The method improves an existing image colorizing method based on local color expansion, reduces severe color permeation problems occurring in a grayscale image colorization process under the condition of less manual line coloring, and improves grayscale image colorization quality.

The invention discloses a gray level image colorization system and method based on a GLCM. The gray level image colorization system comprises an image block module, a GLCM feature extraction module, a similarity matching establishment corresponding relation module, a color marking and preliminary correction module and an optimization coloring module. The deficiencies of previous methods can be overcome, and engineering practice requirements are met. According to the method, the image colorization speed can be increased and the image colorization effect can be improved.

The invention discloses a similar image colorization algorithm based on classification learning. The similar image colorization algorithm comprises the following steps: sample images are collected, an image gradation co-occurrence matrix attribute is extracted, the sample images are classified into five categories through the AP algorithm, superpixels of a target image and superpixels of a reference image are calculated respectively, then, colors are transferred from the reference image to the target image, colors of the superpixels are corrected afterwards according to continuity of image space, and finally the algorithm is used for conducting color diffusion to complete colorization. According to the similar image colorization algorithm, the influence on an image by a global attribute of the image is considered, the image gradation co-occurrence matrix attribute is extracted to conduct classification learning on parameters of a superpixel matching function, as a result, different parametric functions can be provided for superpixel matching on images with different compositions, and the universality of the similar image colorization algorithm on the images is improved; besides, after the matching process, region growing algorithm partition can be conducted at a superpixel level, and color correction can be conducted in a region.

The present invention discloses a black-and-white image colorization process method, including: adopting the watershed segmentation to segment the black-and-white image into a plurality of closed areas; performing area clustering to the segmented closed areas, to obtain target areas for executing colorization process; calibrating the color of each target area to obtain colorization formwork, executing color stuffing to each target area according to the colorization formwork, to obtain an image completing the colorization process. Meanwhile the present invention also discloses a black-and-white image colorization process system. The black-and-white image colorization process method and system of the embodiment of the invention, adopts the watershed algorithm to segment the closed areas and then uses an area clustering method to obtain target areas, stuffing the color corresponding to each target area into each target area according to the calibrated color formwork, thereby implementing the quick and automatic colorization process black-and-white image.

The invention discloses a two-way colorization method for animation images based on a U-shaped periodic consistent confrontation network, relating to the Image processing field, Data collection, pixels of the animation illustration image are unified; A training data set and a test data set are constructed, finally, the capacities of a generator G, a generator F, a discriminating network DX and a discriminating network DY are improved by adopting a cyclic training method with consistent U-shaped periods, a function mapping relation between an image and a color image is found, and bidirectionalconversion of a black-white sketch and a full-color image is realized. According to the method, the features do not need to be extracted manually, a training set does not need to be marked, the workload of a cartoon creator is remarkably reduced, the image colorization and blackening processing efficiency is improved, and great help is provided for the cartoon creator.

The invention provides a natural color night vision realization method based on a single band infrared image. The method is characterized by providing a characteristic vector which is based on multi-scale and spatial context information and used for analyzing pixel points. The method comprises the following steps of: training a characteristic-vector-based natural color model by adopting a sample study method first; and then, establishing the color distribution of an infrared night vision image by using the trained natural color model to realize a process of automatic colorization. The method has the advantages that: the method can work in a single band infrared image wherein a conventional night vision image colorization method cannot work, and the single band infrared image can be endowed with natural color automatically so as to improve the accuracy and the efficiency of target identification and scene understanding. The method can be applied to various types of civil and military systems such as a night driving assisting system of a vehicle, a video monitoring system, a military target tracking and identifying system and the like.

The invention discloses a gray image colorization method based on a convolutional neural network. The gray image colorization method comprises the steps: building the convolutional neural network, andconverting a gray image into a color image, wherein the hidden layer of the convolutional neural network comprises a plurality of connecting layers, and each connecting layer comprises a convolutional layer, a batch standardization layer and a combined nonlinear activation function layer which are connected in sequence; and the combined nonlinear activation function layer comprises a nonlinear activation function layer, a single-channel convolution kernel layer, a batch standardization layer and a normalization layer which are connected in sequence; and the combined nonlinear activation function layer performs nonlinear activation processing on a result after convolution operation in a feature layer-by-feature layer manner. The gray image colorization method has the advantages of automatic colorization, large application scene and the like, and can realize the function by using a small number of layers, and the colorization effect is better than that of the traditional method.

The invention discloses a vehicular infrared image colorization and three-dimensional reconstruction technology which is characterized in that a colorization algorithm based on a random forest classifier and a three-dimensional reconstruction algorithm based on panel parameter estimation are integrated to perform three-dimensional reconstruction on a vehicular infrared image. The method of the invention has the following advantages: an infrared image colorization technology and an infrared image three-dimensional reconstruction technology are integrated, which enables an infrared image to be displayed more visually; the method is applicable to colorization of a variety of vehicular infrared scenes and can obtain a good colorization result; and the method is applicable to changing road scenes.

The invention relates to an image colorization method based on over-segmentation and local and global consistency. The method comprises the following steps of performing initial color marking on a roughly segmented gray scale image; converting the initial color marked image from an RGB color space to a YUV color space characterized by brightness and color component separation; calculating the position of the mark point by using a gray scale histogram; acquiring a semi-automatic color marked image after a color for the mark point is automatically selected; minimizing a colorization framework based on local and global consistency learning to acquire final color components U<^> and V<^>; making an original brightness component Y and the final color components U<^> and V<^> integrated and converted into the RGB space to acquire a final colorized image. The acquired image color is clear and natural. The method has relatively great robustness and stability. A relatively high peak value signal to noise ratio is acquired. The manual interaction complexity is lowered while the image colorization quality is improved. The method can be used in fields of movie and television making, medical image enhancement, advertisement designing, etc.

The invention discloses a gray level image coloring method based on VAE-GAN and a mixed density network. The method comprises the following steps: firstly constructing a VAE-GAN model, converting a color image in a data set into a Lab color space, and obtaining a gray level image; learning ab channel color domain representation of the color image by using VAE-GAN; and secondly, constructing a mixed density network model, and learning mixed Gaussian distribution by taking a gray level image as input and ab channel color domain representation as a label. In practical application, a to-be-detected gray level image is input into the trained mixed density network model, the mixed density network model outputs a corresponding mixed distribution coefficient, a corresponding color domain representation is sampled from the mixed distribution coefficient, and then a decoder of the trained VAE-GAN model is used to decode the color domain representation to obtain a coloring result of the grayscale image. The VAE-GAN and the mixed density network are integrated, so that the image coloring quality is effectively improved.

The invention relates to an automatic near-infrared image colorization model composed of a lightweight image recognition network module and an image colorization CNN module with a fusion layer. Fast-RCNN performs image recognition on near-infrared image, and then collects and selects images similar to the scene from the network, After training with the colorized CNN module with fusion layer, the near-infrared image is input for colorization. Therefore, the colored near-infrared image is output. This method combines the lightweight image recognition model with the image colorization model withfusion layer, which not only colorizes most of the near infrared images, but also improves the problems such as boundary diffuse color, gray and dark color, coloring error and so on in the colorized images.

The invention discloses an image colorization method based on L1 mixed norm solving. The image colorization method includes (1) inputting a gray image, (2) inputting initial color lines, namely painting the color lines on different areas of the input image to change the gray image into a color image with the color lines, (3) inducing U and V channel values for utilization of the Y channel value and converting the color image acquired before into the YUV space, and recording the initial color value of the image in the space as b0, (4) according to the fact that neighboring pixels have the similar color values, performing colorization by diffusing the U and V channel values of the initial color pixels to the surrounding, (5) normalizing the objective function for solving the U channel value to obtain the factor xu of the U channel value and the factor xv of the V channel value, and (6) solving the xu and xv according to the normalized objective function and converting the Y, U and V channel values into the RGB(red, green and blue) space. On the basis of an L1 norm algorithm, the final image is natural in terms of color saturation and has optimal visual effects.

The invention discloses a method for intelligently coloring a microstructure photo shot by an electron microscope and a CNN coloring learner, and the method comprises the following steps of firstly, manufacturing a data set which is a set of a plurality of color SEM images, and learning the data set by a shader; converting a to-be-colored microscopic grayscale photo shot by the electron microscope as an input image into an LAB color space, and intercepting a grayscale image with a pixel size of H*W from the microscopic grayscale photo as an input image XL of a CNN; coloring the shader. According to the present invention, the coloring process of the present invention needs to additionally provide a color image in real life similar to the texture structure of a gray image with a colored target as a reference image, and the end-to-end black box type training is conducted on the convolutional neural network, so that the manual participation is not needed.

The invention discloses a document image processing method and device, and the method comprises the steps: carrying out the image denoising processing of an RGB document image through a negative noisethree-channel color relation, and obtaining a denoised document image; based on the minimum bounding rectangle of each foreground connected region in the denoised document image, obtaining a target rotation direction and a target rotation angle by using a preset correction rule, and performing image correction on the denoised document image to obtain a corrected document image; and for the correction document image, performing image coloring processing and image trimming processing through background three-channel clustering in combination with the RGB document image to obtain a target document image. Visibly, the image processing effect of the RGB document image is improved from the aspects of image denoising, image correction, image coloring and image trimming, and the image processingrequirement of the RGB document image is met.

The invention discloses an image colorization method and system and a computer readable storage medium. The method comprises the following steps: A, converting an image to be colorized from an RGB color space to a YUV color space, and separating Y channel data; B, copying the Y channel data, and constructing two-channel data together with the Y channel data; C, using the data of the two channels as the input of a depth convolution auto-encoder to predict UV channels respectively, wherein the depth convolution auto-encoder is formed by connecting a plurality of rejump layers; D, combining the Ychannel data with the UV channel data predicted in the step C to construct a complete YUV color space image; and E, converting the YUV color space image into an RGB color space image to obtain a final colorized image. According to the method, the problems of model gradient disappearance and over-fitting can be better solved, a better coloring effect and better image definition are achieved, artifacts generated by the image can be effectively reduced, and the color saturation can be effectively enhanced.

Example embodiments may allow for the efficient, edge-preserving filtering, upsampling, or other processing of image data with respect to a reference image. A cost- minimization problem to generate anoutput image from the input array is mapped onto regularly- spaced vertices in a multidimensional vertex space. This mapping is based on an association between pixels of the reference image and the vertices, and between elements of the input array and the pixels of the reference image. The problem is them solved to determine vertex disparity values for each of the vertices. Pixels of the output image can be determined based on determined vertex disparity values for respective one or more vertices associated with each of the pixels. This fast, efficient image processing method can be used to enable edge-preserving image upsampling, image colorization, semantic segmentation of image contents, image filtering or de-noising, or other applications.

The invention discloses a grayscale image coloring method based on a reference image color style. The grayscale image coloring method can be used as a solution for the problems of line draft coloring and black-and-white image coloring. The specific implementation of the method comprises the following steps: 1, drawing up a coloring image category library, collecting a color image data set according to categories, carrying out line draft extraction or graying on the color image data set, and extracting color histogram information of the color image data set; 2, taking the color image, the line draft or the gray level image and the color histogram information as a training set, inputting the training set into the network for training, and obtaining a coloring model corresponding to the category; and 3, re-inputting the image to be colored into the network, and coloring the image by using the model obtained in the step 2; and 4, evaluating a coloring result. The grayscale image coloring method based on the reference image color style can solve the coloring problem of various types of line manuscripts and black-and-white images, and it is proved that the coloring effect is good in practice.

The invention discloses a grayscale image colorization method based on a generative adversarial network, and the method comprises the steps: firstly, selecting a quantitative color image group in a COCO image data set, carrying out the decoloring processing, making a training set, constructing a generative adversarial network architecture, enabling a generator model to complete the pre-training in the generative adversarial network architecture, and carrying out the image colorization. And then alternately training the discriminant model and the pre-trained generative model, adjusting parameters to obtain a trained model, and inputting test data into the model to realize gray level image colorization. Through the pre-training method and process of the generator, the training method and data set optimization are greatly improved, the training time is greatly shortened on the basis of ensuring the training quality and the generalization quality of the finally generated image, and the method has flexibility; and training and testing are carried out on a COCO data set by utilizing a U-Net thought, so that the defects that manual intervention is needed and fine coloring work of a large-size image pixel level is difficult to carry out in a traditional method can be reduced to a great extent.

The invention discloses a multi-scale neural network infrared image colorization method based on an attention mechanism, and the method comprises the steps: carrying out the feature extraction of an input infrared image under different resolution scales through a two-dimensional convolutional neural network, and carrying out the extraction of the extracted high-dimensional feature information through the attention mechanism, and finally carrying out fusion processing on the multi-scale information to obtain a predicted colorized infrared image. Compared with an existing infrared image colorization network, the method has the advantages that a neural network algorithm model is constructed based on an attention mechanism and a multi-scale hierarchical structure, and by adopting an improved spatial attention and multi-dimensional feature connection mechanism, the network model feature extraction capability can be improved while the model complexity is effectively reduced; by designing a composite loss function of pixel loss, edge loss and perception loss, the quality of the colorized infrared image is further improved.

The invention relates to a night vision image colorization method based on guided filtering image fusion, and belongs to the technical field of digital image processing. Aiming at the fusion of an infrared image and a visible light image, various kinds of information are displayed on one fusion image, and a target is highlighted. Specifically, firstly, an image fusion algorithm based on guided filtering is adopted to carry out weighted fusion on two or more multi-band grayscale images, a grayscale fusion image with detail information of multiple images is obtained, then color transmission is carried out on the fusion image according to a Welsh algorithm, and thus a final natural color image is obtained. The color image obtained through the method is superior to a traditional night vision image colorization method in the aspects of image contrast, color brightness, image definition and the like.

The invention provides a safety check image coloring method, a device, a storage medium and the computer equipment. According to the invention, an original high-low-energy image is input into the image coloring model, the ab coloring channel data corresponding to the original high-low-energy image can be obtained, the L channel data based on the original high-low-energy image and the ab coloring channel data output by the image coloring model can be obtained, and the image coloring efficiency is improved. Lab color space data corresponding to the original high and low energy image can be obtained; according to the invention, the image coloring model is used for coloring the original high-low energy image collected during safety check, so that a mode of performing material classification and coloring after calculating a corresponding atomic number through a high-low energy experimental data derivation formula in the prior art is eliminated; the conditions that atomic number calculation is inaccurate and a many-to-one relationship exists between the proton ordinal number and the color space during atomic number calculation are avoided, so that the image coloring accuracy is effectively improved.

In an image processor, images are created, stored, manipulated and regenerated using color tinting, where color tinting applies component-to-color mapping from a color card to a plurality of component images, which are then combined to form a final image, tinted according to the content of the color card. In some instances, the color card might code one color for each of N components, in which case the final image might be the merging of each of N monochromatic component images colored by the color coded by the color card. In other instances, the color card codes for intensity levels or a texture.

The invention discloses an image coloring method based on a multi-residual network and regularization transfer learning. The image coloring method comprises the steps that a gray level image data setis manufactured; extracting image features by using an image feature extraction module constructed based on a multi-residual network; training an image semantic feature extraction module based on a regularization transfer learning framework, and extracting image semantic features by utilizing the image semantic feature extraction module; inputting the image features and the image semantic featuresinto an image fusion module for fusion to obtain fusion features of the grayscale image; and inputting the fusion features of the grayscale image into an image coloring module constructed based on amulti-residual network for coloring to obtain a new color image. According to the invention, the image feature extraction module and the image coloring module are constructed based on the multi-residual network, so that the network performance is improved; an image semantic feature extraction module is trained based on a regularization transfer learning framework, image semantic features are extracted, and the accuracy of semantic feature extraction and the accuracy of image coloring are improved.

The invention discloses a multi-sparse dictionary grayscale map colorization method based on feature classification detail enhancement. Feature classification and a multi-sparse dictionary are combined to establish a colorization processing model, thereby realizing colorization of a grayscale image; a corresponding local constraint algorithm is provided to solve a problem of inaccurate classification, so that the classification accuracy is improved and the colorization effect is enhanced; and then a detailed enhancement algorithm is provided based on a Laplacian pyramid so as to solve a problem of detail losses inherent in sparse representation, so that the detail loss problem is solved and the speed of image colorization is increased. Therefore, the grayscale image colorization effect that conforms to the visual habit of the human being and has the high natural sense is obtained. The method can be applied to other fields like colorization of a grayscale fusion image and an infrared image and color transmission between the color images.

The embodiment of the invention discloses an image processing method and device, electronic equipment and a storage medium. According to the embodiment of the invention, the method includes: acquiringa user line drawing image and a user coloring image specified by a user, and acquiring a to-be-colored target line drawing image, a reference image and a reference line drawing image; coloring the target line drawing image to be colored based on the reference image and the reference line drawing image to obtain an initial target colored image; extracting coloring information of the user coloringimage according to the user line drawing image; and performing coloring adjustment processing on the initial target coloring image based on the coloring information to obtain a final target coloring image. According to the embodiment of the invention, the target line drawing image can be preliminarily colored according to the reference image, then coloring is further adjusted according to the usercoloring image, and finally the target coloring image with the same coloring style as the user coloring image is obtained. Therefore, the scheme can improve the efficiency of the image processing method.

Apparatus for a motor vehicle comprises an image sensor generating non-trichromatic image data as an image pixel array covering a predetermined field of view relative to the vehicle. A color-calibration source generates at least one color reference value according to an object depicted in the non-trichromatic image data. A controller is configured to 1) associate the at least one color reference value to a set of pixels within the image pixel array, and 2) colorize the non-trichromatic image data according to the at least one color reference value to produce a colorized image. A display is configured to display the colorized image to a viewer in the vehicle.

The invention discloses a night image coloring method and device, a medium and equipment, and the method comprises the steps: constructing a coloring network, and training the coloring network to obtain a coloring model; obtaining a visible light image and an infrared image shot at the same visual angle at night; according to a preset fusion rule, performing image fusion processing on the night visible light image and the infrared image to obtain a to-be-colored fusion image; and inputting a fusion image to be colored into the coloring model, and carrying out feature extraction and non-reference coloring. According to the method and device, feature fusion is carried out on the visible light image and the infrared image, the scene detail information in the visible light image and the target feature information in the infrared image are synthesized, and based on the trained coloring model, the non-reference night image coloring method is realized, and the coloring quality of the night image can be effectively improved.

The invention discloses an image coloring method based on an improved depth separable convolutional neural network. The method comprises the following steps: constructing an image data set; constructing an improved deep separable convolutional coloring neural network; training an improved deep separable convolutional coloring neural network; and inputting the grayscale image to be colored into thetrained lightweight colored neural network to obtain an image colorization result. According to the network structure of the invention, global semantic features and local pixel features are comprehensively considered, and residual errors, depth separable convolution, channel weighting and other modes are used to reduce parameters and improve performance.