383 results about "Image compression" patented technology

A system and method for compressing synthetic aperture radar (SAR) images with enhanced phase recovery and unwrapping capabilities is disclosed. The system performs preprocessing on input SAR images, applies discrete cosine transform (DCT) to create subbands, and utilizes a multi-pass amplitude compression technique. A specialized neural network performs phase unwrapping using compressed amplitude information and interferogram wrapped phase data. The system employs a channel-wise transformer fusion block (CTFB) for feature fusion and a multi-stage context recovery subsystem with optimized loss functions for both amplitude and phase recovery. The method achieves improved compression efficiency and phase recovery accuracy, particularly beneficial for Interferometric SAR (InSAR) applications.

Coding an image that includes text content and a background is disclosed. Text portions are identified in the image. The text portions are extracted from the image to obtain a background image, where the background image includes holes corresponding to respective areas of the text portions within the image. A filled-in background image is obtained based on the background image. The filled-in background image is encoded into a compressed bitstream using a block-based encoder. The text portions is also encoded into the compressed bitstream. Encoding the text portions includes encoding respective high quality text binarization upscaled binary maps.

A method implemented by a decoder. The method includes receiving a vision-language control latent feature, a vision-language latent feature of an original image, and a diffusion latent feature of the original image, where the vision-language latent feature comprises text and integers; computing, based on the vision-language latent feature, a decoded vision-language feature; computing, based on the vision-language control latent feature and the decoded vision-language feature, an encoded control feature; reconstructing, based on the encoded control feature and the decoded vision-language feature, a baseline image output; computing, based on the diffusion latent feature, the encoded control feature, and the decoded vision-language feature, a supplementary output; and reconstructing, based on the supplementary output and the baseline image output, a final decoded image output.

The invention discloses a deep learning image compression method and system based on a semantic discriminator, and mainly solves the problem that the visual task performance of a downstream machine is remarkably reduced due to serious semantic information loss under a high compression rate in the conventional image compression method. According to the implementation scheme, a group of images are selected from an existing image data set and are divided into a training set, a verification set and a test set, and the training set, the verification set and the test set are preprocessed respectively; constructing an image compression network comprising an image codec, a semantic extraction network and a semantic guide discriminator under a Pytorch framework; inputting the training set into an image compression network, carrying out two-stage iterative training, and verifying through a verification set; and inputting the test set into the trained image compression network, and only calling the image codec to output the compressed reconstructed image. According to the method, the structural integrity and semantic consistency of the reconstructed image are remarkably improved, higher visual quality and task accuracy can be kept at a low code rate, better image compression performance is embodied, and the method can be used for efficient transmission and storage of image data.

For compressing synthetic aperture radar (SAR) images, preprocessing operations are performed on an input SAR image. A discrete cosine transform is performed on the image, and multiple subbands are created, where each subband represents a particular range of frequencies. The subbands are organized into multiple groups, where the multiple groups comprise a first low frequency group, a second low frequency group, and a high frequency group. A latent space representation is generated corresponding to each of the multiple groups of subbands. A first bitstream is created based on the latent space representation, and an alternate representation of the latent space is used for creating a second bitstream, enabling multiple-pass techniques for SAR image data compression, including phase unwrapping for supporting interferometric SAR (InSAR) applications.

The invention discloses a Mama-based entropy model and an image compression method, and mainly solves the problem of limited compression performance caused by poor entropy estimation precision in the prior art. The scheme comprises the following steps: 1) constructing a two-dimensional state space hyper-prior network by a hyper-prior encoder and a decoder to obtain global hyper-prior features; 2) adopting a double-branch gating architecture to obtain a mixed context feature containing local and global dependency information; 3) constructing an entropy parameter fusion and probability modeling network, carrying out channel splicing and feature fusion on the mixed context features and global super-prior features, and outputting a conditional Gaussian distribution parameter of each latent variable position; according to the method, the entropy model estimation precision can be remarkably improved, meanwhile, the compression rate distortion performance is improved, and the balance between the rate distortion performance and the calculation complexity is achieved.

The invention relates to the technical field of image reconstruction, in particular to an intelligent trolley image compressed sensing reconstruction method and system based on deformable convolution. The method comprises the following steps: an intelligent trolley obtains multiple frames of original images in a driving scene through a vehicle-mounted image acquisition module and carries out preprocessing to generate a preprocessed image set; performing feature extraction, feature matching, hypergraph transformation and image synthesis operation on the basis of the preprocessed image set, performing compressed sensing sampling processing at the same time to obtain compressed sensing sampling data, and inputting the compressed sensing sampling data into a preset compressed sensing reconstruction network; a deformable convolution module is embedded in the front end of the compressed sensing reconstruction network for feature extraction and image detail recovery calculation, and a preliminary reconstruction image is generated; and embedding a deformable deconvolution module at the rear end of the compressed sensing reconstruction network to carry out pixel-level adjustment and quality evaluation until a final reconstruction image meeting the reconstruction effect requirement is generated. According to the invention, the precision of vehicle-mounted image compressed sensing reconstruction can be improved.

The invention discloses a retina fundus image generation method based on a diffusion model. Firstly, a training data set and a regularization data set are constructed; then, constructing a diffusion model which comprises a variational auto-encoder, a U-Net denoising network and a text encoder; the variational auto-encoder comprises a variational encoder and a variational decoder, and the U-Net denoising network is embedded between the variational encoder and the variational decoder; the structured text prompt passes through a text encoder to obtain a text embedding vector, and the vector is injected into the U-Net denoising network as condition information; the variational encoder compresses the retina fundus image into potential features, the U-Net denoising network carries out denoising on the potential features under the guidance of condition information, and the denoised potential features are reconstructed into a high-resolution retina fundus image through the variational encoder; and finally, generating a retina fundus image based on the pre-trained diffusion model. The controllability of image generation is improved, and the generated image is highly consistent between the focus form and the medical description.

The invention belongs to the technical field of image processing, and particularly relates to a remote sensing image compression and reconstruction method based on feature perception and potential diffusion super-division. The method comprises a coding and compression stage and a decoding and reconstruction stage, in the coding and compression stage, intelligent analysis, selective compression and data encapsulation are carried out on an original high-resolution remote sensing image, a feature region and a homogeneous background region in the image are intelligently identified and separated through a convolutional neural network, and differentiation processing is carried out on the feature region and the homogeneous background region; and performing high-precision recording on the feature region, and performing aggressive down-sampling on the homogeneous background region. In the decoding and reconstruction stage, a high-resolution reconstructed image is recovered from a compressed data packet, reconstruction is carried out through a potential diffusion super-resolution model, the model can utilize the semantic context of the image so as to generate visual natural textures, and the problems of overall blur and the like caused by traditional interpolation are effectively avoided. And therefore, the image can still excellently maintain key ground feature details under a high compression ratio.

The invention relates to the cross technical field of image processing and artificial intelligence, in particular to an image compression method combining importance prediction and adaptive sampling compressed sensing, which comprises the following steps: acquiring an image data set, and preprocessing to obtain image blocks; constructing an importance prediction network, analyzing the image blocks through the importance prediction network, and outputting an importance map; carrying out adaptive sampling processing on the importance map, outputting the sampling rate of each image block, and constructing a weighted measurement matrix; performing compressed sensing projection on the image data set based on the weighted measurement matrix to generate a low-dimensional measurement value; and inputting the low-dimensional measurement value into a deep reconstruction network for image reconstruction to obtain a reconstructed image. For relatively high reconstruction delay, invalid measurement and invalid calculation are reduced through cooperative matching of a sampling side and a reconstruction side, and reasoning delay is reduced; the storage and transmission cost is remarkably reduced under the conditions of low sampling rate and low computing resources, and meanwhile, the reconstruction quality equivalent to that of a high sampling rate scheme is kept.

A system processing hardware executes a machine learning (ML) model-based video compression encoder to receive uncompressed video content and corresponding motion compensated video content, compare the uncompressed and motion compensated video content to identify an image space residual, transform the image space residual to a latent space representation of the uncompressed video content, and transform, using a trained image compression ML model, the motion compensated video content to a latent space representation of the motion compensated video content. The ML model-based video compression encoder further encodes the latent space representation of the image space residual to produce an encoded latent residual, encodes, using the trained image compression ML model, the latent space representation of the motion compensated video content to produce an encoded latent video content, and generates, using the encoded latent residual and the encoded latent video content, a compressed video content corresponding to the uncompressed video content.

The invention discloses an image compression reconstruction method and system based on block modulation sequence stream compression. The method comprises the following steps: acquiring an input image and a modulation mask; dividing the modulated image into non-overlapped sub-blocks, and extracting pixels at the same position from each sub-block to form a low-resolution image sequence flow; summing pixels at the same position in all the low-resolution sequence flow images according to index dimensions to generate compression measurement; converting reconstruction into an optimization problem according to a forward measurement model, and constructing an interpretable depth expansion mathematical model; the cascade design of a cross attention enhancement module, a multi-mode expansion Mama module and a channel attention module is combined in a deep denoising approximation operator, and a high-resolution image is efficiently reconstructed through a deep expansion network. The low-complexity forward coding model and the high-efficiency decoder design provided by the invention optimize the coding and decoding efficiency, maintain excellent reconstruction precision while greatly improving the compression ratio, and are suitable for scenes such as a mobile terminal monitoring platform with limited resources.

The invention discloses a diffusion image compression and reconstruction method combining semantic guidance and regional detail enhancement, and the method comprises the steps: extracting image semantic features through a pre-training visual semantic model, carrying out the PCA dimension reduction, generating semantic prior embedding matched with a diffusion potential space, and meanwhile, generating a structure perception ROI mask corresponding to small contents in the potential space; in the diffusion forward coding process, the KL divergence of each time step is adaptively divided between the ROI and the background, and two paths of noise are respectively written in through a Gaussian channel; in a reverse denoising stage, a condition guidance weight of spatial variation is constructed based on semantic prior and ROI mask, and pixel-by-pixel fusion is performed on noise prediction of unconditional branches and conditional branches, so that the definition and the stability of fine structures such as characters and human faces can be remarkably improved under the condition of low bit rate, and the method is suitable for large-scale popularization and application. Meanwhile, the naturalness of the overall structure and texture is kept, and the method has high detail fidelity, good subjective and objective rate distortion performance and high practical popularization value.

The invention discloses a deep learning image compression method and system based on wavelet domain double branches, and the method comprises the steps: carrying out the coding and decoding through a deep neural network in combination with the wavelet domain features of high-frequency and low-frequency double branches, carrying out the coding and mapping of an original image to a compact potential feature, carrying out the super-prior coding, quantization and entropy coding, so as to generate a code stream, and carrying out the compression of a deep learning image. Global context information is obtained through hyper-prior decoding, channel division is performed on the compact potential features, and context modeling based on space and channels is performed on each channel block by using the global context information so as to predict a mean value and a standard deviation of the channel blocks obeying Gaussian distribution, and the mean value and the standard deviation are used for guiding quantization and entropy coding of the channel blocks; and generating a code stream after image compression, mapping the potential features obtained by decoding back to the reconstructed image, and constructing rate distortion loss based on the control code rate of the potential features after decoding, the control code rate of super-prior decoding and the distortion of the original image and the reconstructed image so as to train a deep neural network for image compression.

The invention relates to the technical field of computer graphics and image compression, in particular to a multi-moment illumination chartlet compression and decompression method based on two-dimensional Gaussian representation and a computer device.The method comprises the steps that S1, illumination chartlets at multiple target moments are obtained and preprocessed; s2, constructing a shared two-dimensional Gaussian basis set based on the low-frequency component; s3, extracting a residual error at a multi-target moment and features of a highlight and high-frequency region; s4, multi-layer perceptron network construction and two-dimensional Gaussian attribute offset modeling are carried out; s5, performing compression and storage; and S6, decompressing and rendering. The compression rate is greatly improved, the decompression speed is extremely high, the real-time rendering requirement is met, the rendering quality is higher than that of a traditional compression method, the highlight and high-frequency detail modeling capacity is high, the structure is simple, and integration is easy.

The invention provides an image compression method and device based on cooperation of frequency domain transformation and high-frequency denoising, and the method comprises the steps: a data preparation step: carrying out the cutting and preprocessing of a target optical image, and generating a composite image containing various types and intensities of high-frequency noise for training and testing; a combined image coding and denoising step: setting a double-branch image coding and denoising module which comprises a main branch and a side branch sharing parameters; the main branch takes a noisy image as input, the side branch takes a corresponding clean image as input, the main branch is guided to synchronously learn and denoise in an image feature coding process, and a noiseless image coding feature is output; a frequency domain transformation denoising unit is embedded in the double-branch image coding denoising module; and a feature compression and image restoration step: carrying out quantization and entropy coding on the noiseless image coding features to generate a compressed code stream, and obtaining a final denoised restored image through decoding and image reconstruction.

The present disclosure provides techniques for improving signaling efficiency in the context of rate control governing a desired ratio of bitrate load and quality in image compression. It is provided a method for encoding an image comprising obtaining the image, encoding the image into a bitstream based on a first coding parameter, wherein the first coding parameter is indicative of a first displacement between a first target rate control parameter and a first reference rate control parameter, and encoding the first coding parameter into the bitstream. Further, it is provided a method for decoding an image, comprising receiving a bitstream comprising coded data of the image, parsing the bitstream to obtain a first coding parameter, wherein the first coding parameter is indicative of a first displacement between a first target rate control parameter and a first reference rate control parameter, and reconstructing the image based on the first coding parameter.

The application provides a generative image compression method based on vector quantization, and belongs to the technical field of image and video compression. The method comprises the following steps: obtaining a continuous latent representation of an input image through an analysis transformation module; then performing vector quantization processing to obtain discrete indexes of the continuous latent representation and corresponding quantized features; constructing a continuous index probability distribution; predicting a conditional probability distribution of the discrete indexes through a conditional autoregressive entropy model; calculating a coding rate based on the continuous index probability distribution and the conditional probability distribution; reconstructing an image based on the quantized features and constructing a distortion loss; constructing a rate-distortion loss function and jointly training an image compression model comprising the analysis transformation module, the vector quantization module, the conditional autoregressive entropy model and a synthesis transformation module to obtain a trained image compression model; and compressing the input image by using the trained image compression model. The application can realize rate-distortion joint optimization and collaborative improvement of compression efficiency and reconstruction quality.

According to implementations of the present disclosure, there is provided a context-based image coding solution. According to the solution, a reference image of a target image is obtained. A contextual feature representation is extracted from the reference image, the contextual feature representation characterizing contextual information associated with the target image. Conditional encoding or conditional decoding is performed on the target image based on the contextual feature representation. In this way, the enhancement of the performance is achieved in terms of the reconstruction quality and the compression efficiency.

This invention relates to the field of image processing and data transmission technology, specifically disclosing a method and system for adaptive compression and breakpoint resume scheduling of wildlife images. This invention acquires multi-dimensional data of wildlife images in real time, constructs a quantitative evaluation system, and dynamically optimizes data transmission and processing. First, raw data is collected from the wildlife image acquisition terminal, covering core information such as terminal identification, network status, and image content. Then, value assessment coefficients and breakpoint resume scheduling coefficients are calculated sequentially, and an image compression strategy is obtained accordingly. Adaptive compression and breakpoint resume transmission are then executed. Finally, verification, retransmission, and image enhancement optimization are performed on the server side, achieving closed-loop optimization across the entire link from terminal to server. This ensures high-definition transmission and reliable delivery of rare species images even in weak network environments, improving the efficiency and reliability of the monitoring system.

A method implemented by a decoder includes receiving adjusted masked sparse visual features, adjusted masked sparse textual features, adjusted masked control potential and diffusion potential features of an original image; generating a restored masked visual feature based on the received adjusted masked sparse visual feature; calculating an encoded masked textual feature based on the received adjusted masked sparse textual feature; calculating an encoded masked controlled feature based on the adjusted masked controlled potential, restored sparse visual feature and the encoded masked sparse textual feature; reconstructing a baseline image output based on the restored masked visual features, the encoded masked textual features and the encoded masked control features; calculating a supplemental output based on the diffusion potential feature, the masked sparse visual feature, and the masked sparse textual feature; a final decoded image output is constructed based on the supplemental output and the baseline image output.

The application provides a ball tooth appearance defect optical detection device, which comprises an optical detection system, a plurality of detection stations for placing ball teeth respectively, and top surface detection devices, top end detection devices and a plurality of appearance detection devices corresponding to the plurality of detection stations respectively, wherein the top of the ball tooth is contracted, the top surface detection device comprises a top surface camera for focusing on the outer periphery of the top of the ball tooth to shoot a first top image, the first top image at least comprises a clear ball tooth top outer periphery image, the top end detection device comprises a top end camera for focusing on the top end of the ball tooth to shoot a second top image, the second top image at least comprises a clear ball tooth top end image, and the top end image and the top outer periphery image of the same ball tooth form a complete top image. The segmented shooting mode of the top end combined with the outer periphery makes the top defect detection result not affected by the reflection of the top end and the compression of the outer periphery image, and clear and complete defect detection of the whole top of the ball tooth is stably realized.

The invention discloses a compressed sensing MRI (Magnetic Resonance Imaging) reconstruction method and system based on cross-frame guidance and dual-view interaction enhancement. MRI image pairs are obtained, and image compressed sensing reconstruction is performed by using a network comprising a sampling module, an initial reconstruction module and a plurality of iterative reconstruction modules. Each iterative reconstruction module comprises two branches and a double-view-angle interaction enhancement module, and one branch outputs key frame reconstruction guide information; and the other branch reconstructs the auxiliary frame reconstruction image, and performs adjacent slice joint reconstruction by combining key frame reconstruction guide information. Frequency domain information is reserved by constructing image pairs with different sampling ratios, adjacent slices have high spatial correlation, the cross-frame guide module is used for performing feature interaction and information compensation between the slices with different sampling ratios, the double-view-angle interaction enhancement module is used for enhancing structural continuity from different axial directions, and the structural continuity is enhanced. And the global structure is consistent and details are reserved.

The application discloses a generative image compression training method based on feature space semantic anchor points, comprising: obtaining a training image and mapping it into quantized latent features by using an encoder network; inputting a generative decoding network to obtain a corresponding reconstructed image through a conditional generation process; inputting the training image and the reconstructed image into a pre-constructed and parameter-frozen semantic encoder respectively to extract respective corresponding semantic feature representations; calculating a semantic anchor point loss based on the consistency between the semantic feature representations of the training image and the reconstructed image; and updating the parameters of the encoder network and the generative decoding network accordingly. Through the introduction of a frozen semantic reference and a multi-granularity space alignment mechanism in the feature space, the application effectively solves the semantic drift and spatial structure misplacement problems caused by the generative model in the extremely low code rate compression scene, so that the reconstructed image maintains high perceptual quality while its semantic content and spatial layout are consistent with the original image.

The invention provides an image compression method and system based on a partially reconfigurable and evolutionary CNN. The method comprises the following steps: step 1, acquiring an image data stream; step 2, based on the complexity of the image, carrying out adaptive blocking on the image data stream to obtain image blocks with variable sizes; step 3, inputting the image block into a CNN intra-frame predictor deployed in a reconfigurable partition of the FPGA to generate a prediction block; the CNN intra-frame predictor is a lightweight neural network model subjected to quantization compression; step 4, generating and outputting a compressed code stream based on the image block and the prediction block; and step 5, in response to a trigger condition, in a display frame blanking period, loading the updated partial bit stream to a reconfigurable partition of the FPGA through a partial reconfiguration technology so as to evolve the CNN intra-frame predictor online. According to the method, the block size can be automatically adjusted according to the image content, ARGB image compression of online evolution of a prediction model is supported, and simultaneous rising of compression efficiency, visual quality and scene adaptability is realized.

The present application belongs to the technical field of image processing, and specifically relates to a remote sensing image compression and reconstruction method based on feature perception and latent diffusion super-resolution. The method comprises an encoding and compression stage and a decoding and reconstruction stage. The encoding and compression stage performs intelligent analysis, selective compression and data packaging on the original high-resolution remote sensing image, intelligently identifies and separates the feature region and the homogeneous background region in the image through a convolutional neural network, and differentially processes the two regions, with the feature region being recorded at high precision and the homogeneous background region being aggressively down-sampled. The decoding and reconstruction stage recovers a high-resolution reconstructed image from the compressed data packet, and performs reconstruction through a latent diffusion super-resolution model. The model can utilize the semantic context of the image to generate visually natural textures, effectively avoiding overall blurring and other problems caused by traditional interpolation, and thus enabling the image to still well maintain key feature details under high compression ratio.

A method for encoding image data of an image divided into a plurality of pixel blocks using a machine learning algorithm, is proposed, which comprises, by a computing platform comprising a processor configured for implementing the machine learning algorithm, for a block of the pixel blocks: obtaining a block neighborhood set of at least one pixel of the image, wherein the at least one pixel is located outside the block in a spatial neighborhood of a corresponding pixel located on an edge of the pixel block, and generating, by the machine learning algorithm configured for performing end-to-end image compression, a bitstream representing the encoded pixel block, by encoding the pixel block based on input data comprising the pixel block and the block neighborhood set.

The invention discloses a token stream guide variable code rate image compression method oriented to unification of perception and understanding. The method comprises the following steps of: 1, acquiring a two-dimensional image, and processing the two-dimensional image to acquire a one-dimensional token sequence; 2, processing the one-dimensional token sequence through a variable token mask to generate a binary compressed bit stream; 3, decoding the compressed bit stream to obtain an unmasked token sequence, and carrying out dynamic prediction on the unmasked token sequence to recover a complete token sequence; and 4, realizing human perception through the receiving end I after the complete token sequence is recovered, realizing machine perception through the receiving end II after the complete token sequence is recovered, and finally realizing image compression. The method has the capability that a single model supports the continuous variable bit rate, supports a large language model to directly perform semantic understanding based on the compressed code stream, and has the characteristics of low calculation complexity and low delay.

The invention relates to the technical field of image communication transmission, in particular to a monitoring video compression storage method. The method comprises the following steps: acquiring a frame sequence of a monitoring video and constructing a three-dimensional video tensor; for any coordinate position in any frame of image, calculating the time sequence disorder degree of the coordinate position and the structural consistency of the coordinate position; the adaptive weight of each coordinate position is calculated based on the local structure consistency, an adaptive weight matrix is constructed, and each weight value is inversely proportional to the local structure consistency of the corresponding coordinate position; a weighted low-rank tensor decomposition algorithm is adopted to process the three-dimensional video tensor, a low-rank background tensor and a sparse foreground tensor are obtained, and the adaptive weight matrix is used for applying spatially variable sparsity constraint to the sparse foreground tensor; and compressing and storing the low-rank background tensor and the sparse foreground tensor respectively. The method has the effect of improving the image compression efficiency and fidelity.