17 results about "SRGB" patented technology

The invention provides an image enhancement method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining an intensity component of a to-be-enhanced image in an sRGB color space, and an original hue component and a saturation component of the to-be-enhanced image in an HSV color space; determining a color density modulation factor, a perception hue and a saturation adaptive factor according to the intensity component and the original hue component; determining a color plane component of the image to be enhanced in the HVI color space according to the color density modulation factor, the saturation component, the saturation adaptive factor and the perceptual hue; and inputting the enhanced intensity component, the color density modulation factor, the saturation adaptive factor and the enhanced color plane component into a reverse conversion module to obtain an enhanced image in the sRGB color space, thereby effectively overcoming the inherent defects of the traditional color space in low-light image enhancement.

The invention discloses a closed-loop simulation evaluation method and system for a backlight diffusion model, and relates to the technical field of display systems, and the method comprises the steps: carrying out the calculation based on an initial linear image, and obtaining a partition backlight value; fitting screen front backlight and reference screen front backlight are determined based on the partition backlight values; determining a fitting liquid crystal value based on the fitting screen front backlight and the initial linear image; determining a reference liquid crystal value based on the reference screen front backlight and the initial linear image; multiplying the fitting liquid crystal value and the reference liquid crystal value by the reference screen front backlight to correspondingly obtain a fitting linear image and a reference linear image; respectively converting the fitting linear image and the reference linear image into sRGB formats, and correspondingly obtaining a fitting sRGB image and a reference sRGB image; comparing the fitting sRGB image with the reference sRGB image to obtain a perception difference; and evaluating the image quality of the fitted sRGB image. According to the method, the performance of generating the fast approximation model for fitting the sRGB image can be fully evaluated.

Embodiments of the present application disclose a method for image processing, an electronic device, an apparatus and a computing system, the method comprising: inputting an original RAW image to an image processing neural network; respectively extracting time domain features of the original RAW image by using a time domain UNET subnetwork to obtain a time domain feature map; extracting frequency domain features of the original RAW image by using a frequency domain UNET subnetwork to obtain a frequency domain feature map; and fusing the time domain feature map and the frequency domain feature map to output an sRGB image. In this way, when mapping the original RAW image to the sRGB image, the time domain UNET subnetwork and the frequency domain UNET subnetwork are connected in parallel to respectively extract the time domain feature map and the frequency domain feature map, and the frequency domain feature map can make up for the information defects of the time domain feature map, comprehensively utilize the time domain information and the frequency domain information of the image, further optimize the performance of the image processor, and improve the image processing effect.

The invention discloses a lightweight RAW-to-sRGB image mapping method and system based on single-phase opto-color, and relates to the technical field of computer vision and image processing, and the method comprises the steps: taking RAW image data as input, carrying out preprocessing to suppress noise, generating clean RAW data, and enabling four channels of the RAW image data to correspond to denoised Bayer data; extracting edge features and illumination features from the clean RAW data through an IEEM module, and outputting an illumination attention map from the illumination features; processing the image data into opposite colors, extracting multi-feeling features of opposite color channels, and calculating to obtain single-phase opposite color features; edge constraint is carried out on the single-phase color features by using the edge features to obtain color features after attention constraint and residual fusion; on the basis of color features, an illumination attention map is introduced for pixel-by-pixel weighting, and brightness adaptive adjustment is carried out; and finally generating a feature map. The feature map is processed through a plurality of convolution layers and an up-sampling layer, and the feature map is converted into a final sRGB image.

The application discloses a method for improving color accuracy of a liquid crystal display, which is characterized in that a backlight device of the display is matched, a body color temperature and a luminance characteristic value are designed, a preheating time of the display is set, after the display is started and waits for the preheating time, maximum luminance cycle detection of the whole machine, body color coordinate cycle detection of the whole machine, automatic algorithm debugging of a Gamma multi-stage wide coefficient range, accurate color temperature debugging, sRGB and DCI-P3 color space matrix conversion are carried out in a darkroom environment, then a plurality of color pictures are outputted for color accuracy test, an average color accuracy value is calculated, parameters corresponding to color accuracy values meeting preset requirements are written into the display, and the parameters are called by different color space image modes. The application quantitatively designs liquid crystal backlight luminance, color temperature and other characteristics, then accurately calculates and debugs the display online, corrects color temperature, color gamut overlap, white balance, Gamma and other parameters affecting the color accuracy value, carries out real-time color space conversion according to the measured characteristics of the display screen, adjusts color data and luminance data of each display in real time, so that the color restoration of the liquid crystal display is improved, the color accuracy value of each display is reduced, and the color accuracy is improved.

The invention discloses an sRGB-to-RAW video reconstruction method and system based on compressed key frame guidance, and relates to the technical field of video reconstruction. The method comprises the following steps: designing a key frame selection strategy based on a real-world RAW-sRGB pairwise data set, and dynamically selecting a proper key frame; designing an RAW compression reconstruction module, a time domain alignment module, a space-frequency combined mask module and an sRGB guided refined recovery module, and constructing an sRGB-to-RAW video reconstruction model based on compressed key frame guidance; designing a loss function, training an sRGB-to-RAW video reconstruction model by using a deep learning Pytorch framework, and repeatedly traversing a data set until the model is converged; and inputting the RAW key frame and the sRGB video sequence into the trained sRGB-to-RAW video reconstruction model to obtain a reconstructed RAW video sequence. According to the method disclosed by the invention, the high-quality RAW video is reconstructed under the condition that the storage is limited so as to be used for subsequent video editing or visual tasks.

The application discloses a composite film structure capable of realizing color inkless printing by using ultrafast laser, and a printing method and equipment.The dielectric surface of a metal / dielectric composite film is irradiated by focused ultrafast laser pulses, so that an oxidation reaction is induced on the dielectric film surface to form a transparent oxide film, and finally a three-layer composite film of metal / dielectric / oxide is formed from bottom to top.The thickness of the oxide film at the corresponding position is controlled by changing the total energy density of irradiation, so as to change the color of the composite film surface at the corresponding position.The method for realizing color inkless printing by using ultrafast laser has a resolution of 20000 dpi.The composite film structure can form double-resonance absorption in the visible light band, so that the color gamut formed can cover more than 85% sRGB.The printing method does not involve any organic matter and toxic substances, and is environment-friendly.

Provided in the present application are a color gamut conversion method, an apparatus, a computer device, a computer storage medium and a computer program product. The method comprises: acquiring pixel information and color gamut information of a target image; determining whether the color gamut information meets a condition of the valid color gamut of a display device; if not, on the basis of the pixel information, obtaining chromaticity information corresponding thereto, the chromaticity information comprising chromaticity coordinates and a pixel luminance value; on the basis of a color gamut conversion algorithm and on the basis of the chromaticity information and the valid color gamut, determining target chromaticity information; and sending the target chromaticity information to the display device for display. When the color gamut of a lamp cannot completely present all of colors in the sRGB color gamut, the present application can convert a partial pattern exceeding the sRGB color gamut of a picture into the color gamut of the lamp, thereby ensuring the continuity of overall pattern colors presented by the lamp, and effectively reducing color distortion.

This invention discloses an image detail restoration method based on an attention mechanism in a vehicle-mounted night vision system, comprising: acquiring an image tensor of a low-light night vision image; inputting the image tensor into a pre-trained image generation model and performing feature mapping on it; generating a detail-restored image at the output layer according to the mapping parameters of the feature mapping; wherein, the mapping parameters of the feature mapping represent the color matrix and gamma value of the sRGB standard image learned by the image generation model in pre-training; this invention overcomes the shortcomings of existing image enhancement methods by combining local features extracted by convolutional layers with global dependency features captured by self-attention layers; when processing low-light images, convolutional layers focus on detail restoration, while self-attention layers can capture the relationship between distant pixels in the image, thereby effectively restoring global information, so that the restored image is enhanced in both detail and global structure.

The invention discloses a video denoising algorithm and device based on a dense attention denoising network. According to the method, a four-channel Bayer format RAW noise video is obtained, a sliding window forms a three-frame image group after frame extraction, the three-frame image group is input into an encoder-decoder network containing a triple attention module for denoising, and an sRGB video is generated through image signal processing. The method has the advantages of network lightweight design, low parameter quantity, denoising precision and real-time performance, solves the problems of sRGB domain processing defects and difficult deployment of a traditional model, and is suitable for an embedded platform.

The application discloses a low-illumination image enhancement modeling method and device, an enhancement method and a storage medium, and relates to the field of image enhancement. The method comprises training a low-illumination image enhancement model by using training images, and the training process comprises the following steps: inputting the training images into a color gamut conversion module to obtain a linear rawRGB image and a reconstructed sRGB image; inputting the linear rawRGB image into a down-sampling module to obtain a first sub-image and a second sub-image; performing enhancement processing on the second sub-image; extracting a frequency domain brightness guide signal of the first sub-image and the enhanced second sub-image to generate a guide signal; based on the guide signal, using an illumination enhancement decoupling network to estimate a reflectivity component, an illumination component and a brightness enhancement parameter in parallel; and calculating a multi-task loss function based on a model output result and updating model parameters. The application solves the dependence on paired training data which is difficult to obtain, and maintains excellent brightness improvement and detail restoration effects.

The invention relates to the technical field of light control, and discloses a color mapping method and device, a light display device and an electrical equipment, and the method comprises the steps: obtaining the coordinate information of a target point selected by a user on a palette, the palette being formed by setting a canvas containing a plurality of colors; determining color space information of the target point on the palette according to the coordinate information; and determining target RGB intensity according to the color space information, so as to drive a light-emitting device to map a target color corresponding to the target point based on the target RGB intensity. According to the invention, the precise palette is constructed based on the iridescent color ring canvas, and the HSV driving scheme and the sRGB driving scheme are subjected to advantage combination on the basis, so that the color presented by the light-emitting device and the color selected by the palette can be ensured to correspond to actual light, and the consistency of color selection and display is ensured; the natural spectrum color can be continuously and accurately restored on light-emitting devices such as LED lamp balls, the color requirements under multiple scenes are met, and the user experience feeling and the satisfaction degree are improved.

The invention discloses an image encryption and lossless recovery method based on fluorescence chromaticity basis decomposition, which relates to the technical field of digital image processing and information security, and comprises the following steps: constructing a basis matrix based on CIE xy chromaticity coordinates of at least three samples, converting an input image from sRGB into XYZ, and performing non-negative linear decomposition on each pixel to obtain a weight; quantizing the weight, encrypting a quantization result, calculating a residual error of a reconstruction result subjected to inverse quantization after quantization, encrypting and storing according to needs, decrypting and inversely quantizing the weight at a decryption end, reconstructing a lossy image based on the weight, and recovering an original image in combination with the residual error; in the whole process, firstly, a spectrum base is used as a bottom layer representation of pixel decomposition, image coding and physical spectrum / material characteristics are coupled, so that a coding result can be physically mapped to a specific material or spectral response, a residual error is calculated by using reconstruction after inverse quantization at an encryption end, the consistency of quantization and rounding is ensured, and the coding accuracy is improved. Therefore, precise and lossless recovery is realized when the residual error is stored.

The invention provides a color gamut conversion method and device and computer equipment. The method comprises the following steps: acquiring pixel information and color gamut information of a target image; judging whether the color gamut information meets the condition of an effective color gamut of the display equipment or not; if not, chromaticity information corresponding to the pixel information is obtained according to the pixel information, and the chromaticity information comprises chromaticity coordinates and pixel brightness values; determining target chromaticity information according to the chromaticity information and the effective color gamut based on a color gamut conversion algorithm; and sending the target chromaticity information to the display device for display. Under the condition that the color gamut of the lamp cannot completely present all colors in the sRGB color gamut, the partial pattern exceeding the sRGB color gamut of the picture is converted into the color gamut of the lamp, the continuity of the color of the whole pattern presented by the lamp is ensured, and the color distortion phenomenon is effectively reduced.

The present application belongs to the technical field of image processing, and specifically relates to a real-time low-illumination image enhancement method based on a convolutional neural network. The method comprises the following steps: (1) preprocessing a low-illumination RAW image, including rearrangement, normalization and pre-amplification; (2) constructing a network ADU-Net for low-illumination image enhancement, wherein a feature extraction module based on a hollow convolution and a residual connection is used, and an adaptive feature fusion module based on attention is used; (3) inputting the low-illumination RAW image into the ADU-Net to obtain an enhanced sRGB image. With the high efficiency and light weight of the ADU-Net, the present application can restore the low-illumination image at a speed close to real time, effectively improve the brightness of the low-illumination image, accurately restore the color and detail information of the image, and obtain a satisfactory visual effect.

Embodiments of the present application provide an image enhancement method, a chip, and an image acquisition device. The method comprises: storing, by a main processor, a preset number of Raw image frames to be processed currently acquired from a sensor in a first off-chip memory, into the on-chip memory; performing, by a first neural network processor, image fusion on a preset number of Raw image frames to be fused currently obtained based on the preset number of Raw image frames to be processed currently in the on-chip memory based on a preset image fusion network, to obtain a fused Raw image frame, and storing it into the on-chip memory; sending, by the main processor to an ISP chip, a target Raw image frame to be sent currently obtained based on the fused Raw image frame in the on-chip memory. Based on the above processing, image fusion can be performed on the Raw image frames sent by the sensor, during which noise in the Raw image frames can be removed, thereby improving the image quality of the sRGB image frames obtained by the ISP chip.