219 results about "Image code" patented technology

The invention belongs to the technical field of computer vision and deep learning, and particularly relates to an optical flow estimation method and system fusing Mama and visual basic model knowledge. The method comprises the following steps: performing down-sampling feature extraction on two adjacent frames of input images by using a convolutional neural network to obtain local texture features; performing down-sampling on the first frame image to obtain context features; meanwhile, extracting global semantic features of two adjacent frames of images by using a pre-trained visual model, and performing adaptive fusion enhancement through an adaptive semantic texture feature fusion module to obtain an image coding feature pair after semantic enhancement; constructing a related volume through pixel-by-pixel dot product operation; and finally, based on the obtained related volume and context features, iteratively optimizing the output optical flow through a loop iteration updating module. The method solves the problems that in a low-texture, repeated-texture or sheltered area, feature expression is unstable, self-adaptive modeling capacity is lacked, different scenes are difficult to generalize, and model performance and efficiency cannot be balanced.

The invention discloses a video generation method based on a body mask, belongs to the technical field of artificial intelligence, and can solve the problems that an existing body world model is inconsistent in an action space and a pixel space, is sensitive to the change of a visual angle of a camera, and is not uniform in architecture among different body structures. The method comprises the following steps: S1, determining a mask sequence with a body according to a target video; s2, encoding the body mask sequence and the initial frame image of the target video respectively to correspondingly obtain body mask features and image encoding features; s3, inputting the body mask features into a control network module to obtain injection features, and inputting the image coding features into a backbone network of a video generation model to obtain backbone layer features; and S4, fusing the injection feature and the trunk layer feature to obtain a fused feature, and generating a prediction video according to the fused feature. The method is used for generating the predictive video with the body.

The invention provides a road extraction method in order to solve the problem that extracted roads are not connected due to the fact that an existing large-scale road extraction method based on remote sensing images is insufficient in discontinuous road structure modeling capacity and lack of display topology supervision. According to the invention, a model construction method based on a wave particle dipictorial view angle is adopted, particle features and volatility features in an image are extracted respectively, the volatility features output by a fluctuation encoder and the particle features output by a particle encoder are fused through a feature interaction module, the road structure detail and semantic expression ability is enhanced, and the image quality is improved. The method solves the problem that the road features are submerged or neglected in the extraction stage, and achieves the screening and full storage of image coding information. A geometric perception multi-constraint loss function designed for road topological structure features is adopted to perform supervised learning training on the node extraction network, so that the problem of weak supervision pertinence of the node extraction network in a training process in the road extraction network is effectively solved, and the accuracy of road extraction is improved.

The invention discloses an ultrasonic image boundary perception segmentation method, system and device, and relates to the field of image processing, and the method comprises the steps: obtaining an ultrasonic image, and sequentially processing the ultrasonic image through a block embedding module and an image coding module to obtain spatial domain features; performing frequency domain feature extraction processing on the spatial domain feature by using a preset multi-scale frequency extraction strategy to obtain a fine-grained high-frequency feature, a coarse-grained high-frequency feature and a low-frequency feature, and obtaining a frequency fusion feature in combination with a preset frequency alignment strategy; a preset frequency guide boundary refining strategy is combined to determine refining features; and according to the refined features and a preset boundary guiding decoding strategy, determining a segmentation mask representing a result of segmenting the target region under the ultrasonic image from the background region. A low-frequency structure and multi-scale high-frequency boundary details are explicitly separated through frequency domain decomposition, an ultrasonic image boundary sensing segmentation scheme based on frequency guidance is provided, the boundary sensing ability is high, and the cross-domain generalization ability is high.

The invention provides a multi-source solid waste comprehensive road system and method for road reconstruction and expansion, and relates to the technical field of multi-source solid waste distribution. The system comprises a solid waste classification and coding module, a standard solid waste three-level classification system is established, and a unique digital identity code is generated for each kind of solid waste; the multi-source solid waste database construction module is used for storing and managing digital identity codes, attribute data, image information and geographic position information of the solid wastes and constructing a multi-source solid waste database; the solid waste image recognition and search module determines a solid waste image code and inputs the solid waste image code into a multi-source solid waste database for search matching to obtain a target solid waste attribute file corresponding to the solid waste image code; the intelligent matching module for the solid waste road receives project requirements of road reconstruction and extension, performs multi-target optimization matching on the project requirements and the target solid waste attribute file through a built-in intelligent algorithm, and outputs a recommended solid waste recycling scheme; and optimal utilization of solid wastes in road use is selected, so that high value of resources is realized.

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.

According to an embodiment (embodiments) of the present document, multilayer-based coding may be executed. In the multilayer-based coding, a decoded picture based on a target output layer can be output, and, for example, the target output layer can be determined on the basis of an external means or signaled information. According to an embodiment according to the present document, image information acquired from a bitstream can include a specific NAL unit for scheduling target output layer-related information.

The invention discloses a laser marking quality automatic evaluation method based on machine vision. The method comprises the following steps of S1, collecting a laser marking image and performing preprocessing; s2, inputting the preprocessed image into an image segmentation network constructed based on FastSAM to generate a segmentation mask image; s3, extracting a marking area communication block according to the segmented mask image, and constructing an image sub-area set; s4, normalizing the sizes of the image sub-regions, constructing three-channel enhanced input, inputting the three-channel enhanced input into an image coding network constructed based on the OfficientViT, and extracting an image-level feature vector; s5, inputting the image-level feature vector into an ArcFace angle interval classification module, and outputting a defect type classification result; s6, calculating a multi-dimensional quality scoring index; and S7, inputting the multi-dimensional quality scoring indexes into a weighting function to generate a comprehensive scoring result, comparing the comprehensive scoring result with a qualified threshold, and outputting an evaluation result. According to the method, a region perception image segmentation strategy and a lightweight feature coding mechanism are combined, and automatic identification and quality evaluation of character defects in the laser marking image are realized.

The present disclosure provides an improved decoding method and improved method of encoding a bitstream to ensure backwards compatibility of bitstreams with legacy decoders. The decoding method (1400) comprises: receiving (1401) a bitstream including encoded data of an input signal, wherein the bitstream is formed of multiple substreams, a first substream comprising a parameter indicating a number of channels in a second substream for a component of an image coding format of the bitstream; parsing (1402) the first substream, the parsing of the first substream comprising parsing syntax elements of the parameter to thereby determine the number of channels in the second substream for the component of the image coding format; and parsing (1403) the second substream in dependence on the determined number of channels in the second substream for the component of the image coding format. This may allow the number of channels to be decoded in the second bitstream to be signaled in the first bitstream and allow the number of channels used for reconstruction to be reduced.

The application provides a pathological section image analysis method and device based on an image coding large model, and belongs to the technical field of medical images. The method comprises the following steps: cutting an original pathological section image to obtain a pathological section sub-image set; sequentially inputting the pathological section sub-images in the set into a target region segmentation large model to obtain corresponding local target region distribution maps, and then obtaining an overall target region distribution map and generating a set of positive local target region distribution maps; based on the set of positive local target region distribution maps, using an abnormal lesion detection network to generate an abnormal lesion distribution map; and based on the target region distribution map and the abnormal lesion distribution map, finally realizing statistical analysis on the original pathological section image. The application can analyze various pathological section images, automatically identify and classify abnormal lesion regions in the images, and automatically statistically analyze to obtain key clinical indicators such as lesion rates, thereby improving the accuracy and efficiency of diagnosis.

The invention discloses a coke residue feature detection method and system based on image and pressure alignment, and solves the problems that the subjectivity is high, the accuracy is difficult to guarantee and the physical and psychological health of detection personnel is influenced when coke residue feature detection is manually carried out in the prior art. The method comprises the steps of collecting a coke residue image and pressure data, performing preprocessing to form image input and pressure input, and constructing sample data; a neural network model is constructed, image features of image input are extracted by the image coding layer, pressure features of pressure input are extracted by the pressure coding layer, and the image features and the pressure features are fused and then are subjected to coke residue classification through a classifier; constructing a model comprehensive loss function; inputting sample data for training to obtain a coke residue classification model; and deploying the model, and inputting the to-be-detected coke residue image and the pressure data into the model to obtain a coke residue classification result. According to the method, the neural network is fully utilized, the image and the pressure are effectively fused, information complementation and integration are realized, and the classification detection precision of the coke residues is effectively improved.

To provide a method for increasing image / video coding efficiency.SOLUTION: According to the embodiments of the present document, etc., sub-pictures and / or virtual boundaries can be used for coding an image. For example, sub-pictures in the current picture can be used for predicting, reconstructing, and / or filtering the current picture. Virtual boundaries can be used for filtering reconstructed samples of the current picture. Through image coding based on the sub-pictures and / or virtual boundaries according to the embodiments of the present document, etc., the subjective / objective quality of an image can be improved, and the consumption of hardware resources necessary for the coding can be reduced.SELECTED DRAWING: Figure 5

The application discloses a kind of based on image coding technique's transformer area load component identification method, this method is first by analyzing the cluster characteristics of different types of load components, filters and determines the characteristic power curve of each component, and power curve is encoded into two-dimensional image using gram angle field method;Then, based on the measured data of a small number of users, using the minority class oversampling technique to generate the similar power curve of each load component, and then generate the training data set of load component identification model;On this basis, input load component image, use the model training using synthetic data set, identify the power ratio of each load component in transformer area by the model after training. The present application does not need meteorological, environmental and other external factors data, under the actual measurement of current user power data, effectively improve the identification accuracy of transformer area load component, provide accurate information support for load forecasting, demand response and other measures of distribution network.

The invention discloses an image filtering method and device, a storage medium and electronic equipment, and relates to the technical field of image processing, and the method comprises the steps: obtaining a sampling range corresponding to a to-be-filtered pixel point; determining at least one target pixel point adjacent to the to-be-filtered pixel point from the sampling range; determining a first filter coefficient corresponding to each target pixel point and second filter coefficients corresponding to the remaining pixel points except at least one target pixel point in the sampling range; and filtering the pixel point to be filtered based on the first filtering coefficient and the second filtering coefficient. According to the invention, the precision of the plurality of filtering coefficients in the acquisition range corresponding to the target pixel point can be improved, the processing precision of the adjacent region of the to-be-filtered pixel point can be improved, the overall detail reconstruction effect of the image can be effectively improved, and the control of the image code rate can be considered on the basis of improving the precision.

The invention relates to a rolling bearing fault diagnosis method based on multi-sensor fusion and a GAF-CNN. The method comprises the steps of collecting a plurality of one-dimensional time sequence signals of a rolling bearing through a plurality of sensors; converting each one-dimensional time sequence signal into a corresponding two-dimensional time sequence diagram by adopting a Grubrum angle field image coding algorithm; inputting the plurality of generated two-dimensional sequence diagrams into a pre-established neural network model, and extracting a feature vector from each two-dimensional sequence diagram in parallel by the model so as to output a plurality of feature vectors; performing feature fusion on the plurality of output feature vectors to generate a fusion feature; and finally carrying out fault diagnosis on the rolling bearing based on the fusion feature. The invention aims to solve the technical problems of low diagnosis reliability caused by incomplete information of a single sensor and insufficient diagnosis precision caused by insufficient feature extraction when a one-dimensional time sequence signal is directly input into a neural network in the prior art.

This invention discloses a method and system for generating digital watermarked images. The method includes: acquiring an original image to be watermarked; generating an image coding mask based on the image complexity of the original image; performing frequency domain decomposition on the original image to obtain a low-frequency component image; embedding an initial watermark image into the low-frequency component image according to the embedding strength values ​​corresponding to the masks in the image coding mask to obtain a low-frequency watermark image; different masks in the image coding mask correspond to different embedding strength values; and performing an inverse frequency domain transform on the low-frequency watermark image to obtain a synthesized watermark image. This invention can improve the success rate and accuracy of watermark extraction.

The application provides an encoding method, a decoding method and related devices. The encoding method comprises: performing subgraph division on a to-be-encoded image to obtain N subgraphs. For each subgraph in the N subgraphs, the following operations are performed: performing wavelet transform on the subgraph to obtain wavelet coefficients of a low-frequency subband and wavelet coefficients of a high-frequency subband of the subgraph; obtaining first encoding data of the subgraph based on the wavelet coefficients of the low-frequency subband of the subgraph; obtaining second encoding data of the subgraph based on the wavelet coefficients of the high-frequency subband of the subgraph; and obtaining an image code stream of the to-be-encoded image according to the first encoding data and the second encoding data. In the application, the image is encoded or decoded based on a wavelet transform architecture and with a subgraph as a granularity, so that the encoding and decoding efficiency is effectively improved while the encoding and decoding complexity is reduced.

The image decoding method includes determining a context for use in a current block to be processed, from among a plurality of contexts, wherein in the determining: the context is determined under a condition that control parameters of a left block and an upper block are used, when the signal type is a first type; and the context is determined under a third condition that the control parameter of the upper block is not used and a hierarchical depth of a data unit to which the control parameter of the current block belongs is used, when the signal type is a third type, and the third type is one or more of (i) “merge_flag”, (ii) “ref_idx_l0” or “ref_idx_l1”, (iii) “inter_pred_flag”, (iv) “mvd_l0” or “mvd_l1”, (v) “intra_chroma_pred_mode”, (vi) “cbf_luma”, and (vii) “cbf_cb” or “cbf_cr”.

The application discloses a zero-shot abnormal image detection method based on a learnable prompt. A learnable prompt generation module based on context optimization is designed, which contains a learnable prompt and an image abnormal state prompt that can be optimized. A multi-level visual coding feature of a to-be-detected image is obtained by using an image coding network of a visual language large model, and a text feature of a learnable prompt embedding is obtained by using a text coding network. A multi-level cosine similarity between the visual coding feature and the text feature is calculated to construct an image abnormal area calculation module, so that an abnormal area of the to-be-detected image is obtained. The learnable prompt avoids the complexity and instability of manually designed prompts, improves the accuracy of image abnormal detection, guarantees the effectiveness and efficiency of zero-shot learning, and greatly reduces the cost of pre-training of a visual language large model to a downstream task.

The application provides an encoding method, a decoding method and related devices, the method comprising: obtaining to-be-encoded data of an original image, the to-be-encoded data comprising quantization coefficients and control information; and obtaining an image code stream of the original image based on the to-be-encoded data. The image code stream comprises arithmetic encoding data and VLC encoding data. The arithmetic encoding data comprises quantization coefficients less than a first value and control information, and is obtained by using an arithmetic encoding mode for entropy encoding. The VLC encoding data comprises quantization coefficients greater than or equal to the first value, and is obtained by using a VLC entropy encoding mode for entropy encoding. The control information is used for decoding the image code stream. In the application, the characteristics of the VLC entropy encoding mode and the arithmetic encoding mode are combined to effectively reduce the encoding and decoding complexity.

To provide a method, process and system for integrating reconstruction within a next generation video codec for encoding and decoding images given a sequence of images in a first codeword representation, as well as a syntax method for signaling reconstruction parameters and an image encoding method optimized with respect to reconstruction.SOLUTION: In an encoder 200D_E with hybrid in-loop reshaping allowing to encode a part of an image with a second codeword representation allowing more efficient compression than with a first codeword representation, intra slices are encoded according to an in-loop intra reshaping coding architecture and an intra / inter slice switch allows to switch between the two architectures depending on the slice type to be encoded.SELECTED DRAWING: Figure 2G

The application discloses a kind of based on third-generation sequencing genome short variation deep learning detection method and system, by setting the image coding mode of genome sequence generated to third-generation sequencing platform, and according to real variation set and corresponding sequence alignment data establish training set, verification set and test set;Convolutional neural network and the deep learning multi-task classifier integrated by bidirectional long short-term memory neural network is constructed, training set and verification set are used to train and verify deep learning classifier, and the accuracy of deep learning classifier is tested using test set;Based on the deep learning classifier trained, the classification prediction of the stacked image generated by sequence alignment or real variation set is carried out;According to the classification prediction result of stacked image, variation site detection is carried out to sequence alignment data, and complete candidate variation information is obtained, to realize the automatic detection of genome SNP and INDEL short variation.

This invention relates to methods for encoding or decoding video data and methods for storing bitstreams. Specifically, it relates to efficiently signaling the intra-prediction mode used to predict the current block during intra-frame predictive coding. According to one aspect of the invention, an image coding apparatus divides a plurality of intra-frame modes into multiple groups and selects the group to which the actual intra-frame mode of the current block to be encoded belongs, and the image coding apparatus signals the value corresponding to that group. An image decoding apparatus obtains information from the bitstream about the group to which the actual intra-frame mode of the current block belongs, and then selects the final intra-frame mode by evaluating the intra-frame modes belonging to that group.

The invention relates to the technical field of image processing and data compression, and discloses a multi-scale semantic and edge prior fused extensible image coding system and method, which comprises the following steps of: firstly, extracting an edge feature map and a semantic feature map from an original image, and performing compression by using an encoder to obtain an edge feature map and a semantic feature map; generating a feature coding layer code stream oriented to machine vision analysis; secondly, potential representation of an original image is extracted by analyzing a transformation network, and entropy coding and entropy decoding are carried out by using a hyper-prior model and a slice-level autoregressive network; and finally, in the synthesis transformation network of the decoding end, the decoded edge and semantic features are efficiently fused into a main decoding path of the synthesis transformation network on different resolution levels through a multi-scale feature fusion network, and high-quality image reconstruction is realized. According to the method, feature redundancy is suppressed through grouping space attention, gating reconstruction and channel fusion, and the perceptual fidelity and the structural consistency of the reconstructed image under the low bit rate can be remarkably improved.

The present disclosure belongs to the field of image processing technologies, and in particular, to an image decoding method and apparatus, an image coding method and apparatus, a device and a storage medium. The image decoding method of the present disclosure includes: extracting image residual data or extended residual data from an image bitstream, and obtaining a plurality of extended residual groups based on the extracted image residual data or extended residual data; obtaining respective image reconstruction features corresponding to the extended residual groups by performing residual restoration on each of the plurality of extended residual groups; obtaining reconstructed feature data by performing spatial resolution amplification processing on the respective image reconstruction features corresponding to the extended residual groups; and obtaining a reconstructed image block by performing image reconstruction according to the reconstructed feature data.