44results about How to "Improve segmentation" patented technology

The application discloses a kind of modal missing medical image segmentation method, device, terminal and storage medium, belong to medical image processing technical field, including obtaining three-dimensional medical image, the three-dimensional medical image regularization obtains the three-dimensional medical image after regularization;The three-dimensional medical image after regularization is extracted and integrated by encoder to obtain high-dimensional image feature output data;The high-dimensional image feature output data is further enriched by multistage knowledge distillation method and obtains high-dimensional image feature;The further enriched high-dimensional image feature is input into segmentation decoder to obtain final segmentation result.The application extracts more potential information from single mode input, finally improves the segmentation performance in single mode situation, and obtains high-quality segmentation result.

This invention relates to a semi-supervised segmentation method for thyroid nodules based on morphological progressive pseudo-supervised refinement. Addressing issues such as blurred boundaries, background interference, and low pseudo-label quality in semi-supervised learning within ultrasound images, this invention proposes a plug-and-play semi-supervised segmentation module. This module progressively filters blurred boundary responses, recovers reliable foreground regions, and introduces morphologically-aware geometric constraints to systematically refine noisy pseudo-supervised signals. The semi-supervised segmentation module constructed by this method includes: a boundary and uncertainty curriculum module to suppress blurred boundary responses; a dynamic region expansion module to recover weak but reliable foreground regions under spatial constraints; and a gradient-driven geometric refinement module to further standardize boundary quality and structural consistency through topology-aware feedback and differentiable geometric refinement. Through the collaboration of these three modules, the semi-supervised segmentation method gradually transforms noisy pseudo-labels into more reliable and structurally sound supervised signals.

The application relates to the field of sewage pumps, in particular to a novel flow control submersible sewage pump, which comprises a pump body, a rotating shaft arranged in the pump body and an impeller fixedly installed on the rotating shaft, and the bottom end of the pump body is fixedly connected with a sealing cylinder; the bottom end of the sealing cylinder is provided with a screening and dividing assembly, the screening and dividing assembly is used for screening and preliminary dividing of garbage in sewage, and the screening and dividing assembly comprises a screening cylinder fixedly connected with the bottom end of the sealing cylinder, and a dividing cutter capable of moving up and down to cut is arranged in the screening cylinder. The screening and dividing assembly is arranged, so that when the sewage pump is used, the screening cylinder can screen garbage, garbage with a larger size cannot enter the screening cylinder, garbage that can enter the screening cylinder can stay in the screening cylinder for a short time, the garbage is preliminarily divided by the rotating dividing cutter, the size of the garbage is reduced, the pressure of subsequent crushing treatment of the garbage is reduced, and the garbage is conveniently discharged subsequently.

This application provides an open-vocabulary remote sensing image semantic segmentation method and system. The method acquires remote sensing images and learnable text, and obtains image-encoded features and text-encoded features through encoding operations. Low-frequency components are extracted from the image-encoded features, and structured noise is injected into these features to obtain low-frequency enhancement features. A similarity cost map is generated by calculating the similarity between the text-encoded features and the low-frequency enhancement features, and a semantic score is formed by aggregating these similarity cost maps. Detail-guided features are acquired, and the semantic score and detail-guided features are combined to output the segmentation result through image decoding. This method can analyze the low-frequency dominance of remote sensing images based on their frequency composition and enhance discriminative ability in the frequency domain through noise injection, thereby improving the segmentation performance of remote sensing images under open-vocabulary settings and increasing the recognition efficiency of remote sensing images.

The invention discloses a rock core CT image crack segmentation method and device based on deep learning. The method comprises the following steps: making a training data set; expanding the number of training samples in the training data set to form a new data set; a deep learning network model is constructed, the deep learning network model comprises a U-Net network, a contraction path for feature extraction in the U-Net network adopts a feature extraction network based on a VGG-16 network, and an up-sampling expansion path in the U-Net network is added with a CBMA attention module. The CBMA attention module is used for changing the attention of the model to different spatial positions and different channels in the feature map and improving the weight of the area where the rock core crack is located; performing training and parameter adjustment on the deep learning network model by using the training set until the model converges; and testing the trained deep learning network model by using the test set, and evaluating the crack segmentation effect of the deep learning network model. According to the invention, the efficiency of rock core CT image automatic analysis work can be improved.

The application provides a novel feature up-sampling method for semantic segmentation, which improves the resolution of an image by combining pixel reorganization Pixel Shuffle with an attention mechanism, prevents a chessboard effect, and avoids a 0 filling operation on unknown pixel points compared with a previous traditional up-sampling module; secondly, the attention mechanism is introduced to combine global information with local information, obtain a feature map with a larger receptive field and more detailed information, increase effective feature information, and realize the function of effectively improving the segmentation performance of the network. The application has universality, eliminates the limitations of previous methods by simply replacing all up-sampling modules, obtains more representative feature mappings, obtains an enlarged image with good features, makes the restored image obtain more feature information, and is beneficial to improving the segmentation result of the network model.

This invention discloses a boundary enhancement Mamba framework for land-sea segmentation in optical remote sensing and its application. The framework includes: a GCMamba module for extracting multi-scale global-local enhanced feature maps from the input image; a frequency domain boundary enhancement module (FBEM) connected to the output of the global-local features extracted by the GCMamba module, which enhances the features of the land-sea transition region by jointly modeling with frequency and spatial domain information, and outputs a boundary enhancement feature map of the same resolution; and a multi-level feature aggregation pyramid (MFAP), which adjusts the multi-scale input feature maps enhanced by FBEM to a uniform size and number of channels before stitching them together to obtain global features, calculates the correlation between each local feature and the global feature to generate an attention mask to obtain refined features, and finally fuses the refined features with the upsampled features of the previous layer layer by layer to calculate the predicted land-sea segmentation map, thereby effectively solving the problem of blurred land-sea segmentation boundaries in complex environments and improving the accuracy of land-sea segmentation boundaries.

The application is a medical image segmentation method based on multi-modal self-supervision. First, the multi-modal medical image of the lesion tissue is obtained, including A-mode image and B-mode image, and the image is preprocessed. Then, a cycle-consistent modal contrast domain translation network is constructed, including two generators and two discriminators. The generator is used to convert the image of one mode into the image of another mode, including an encoder, an intermediate shared module and a decoder. The discriminator is used to judge the source of the input. Then, the cycle-consistent modal contrast domain translation network is pre-trained, the training loss is calculated, and the loss function includes multi-modal semantic consistency loss, adversarial loss, cross-domain translation loss and cycle consistency loss. Finally, the A-mode segmentation network and the B-mode segmentation network are constructed, the pre-trained weights are migrated to the two segmentation networks, and the trained two segmentation networks are respectively used for medical image segmentation of the corresponding mode. The contrast cross-domain translation is used as a multi-modal self-supervised pre-training task to learn more comprehensive modal features, promote the network to better learn modal characteristics and common knowledge, and improve the segmentation ability.

The invention relates to a semi-supervised semantic segmentation method based on adaptive pseudo label generation, which realizes adaptive control of pseudo label quality by constructing category prototype representation containing a feature mean value and a standard deviation and introducing standard deviation information into a pseudo label screening and prototype consistency learning process. The method comprises a class prototype generation module, an adaptive pseudo-label generation module and a prototype consistency learning module, dynamically adjusts a pseudo-label screening threshold by combining prediction probability distribution and class prototype similarity, and introduces a prototype-based feature consistency constraint in a training process. Therefore, the segmentation precision and the training stability of the semantic segmentation model are remarkably improved under the condition of a small amount of annotated data.

The application discloses a kind of automatic prompting ultrasound image segmentation method and system based on SAM, the method of the present application includes using ultrasound image segmentation network model to carry out ultrasound image segmentation, including CNN encoder, classification head module, low-level feature refining module, SAM's Transform encoder, cross-branch attention module, high-level feature enhancement module, pixel decoder, prompt encoder and mask decoder, pixel decoder decodes and generates preliminary segmentation prediction map, and mask decoder is based on the point, mask embedding vector generated by prompt encoder and the image embedding obtained by CNN encoder and Transform encoder Decoding obtains the final segmentation prediction map.The present application aims to solve the problem that most of the existing segmentation methods based on SAM rely on artificial prompting and have poor adaptability to ultrasound images, reduce the false segmentation affected by ultrasound noise and shadow, and improve the segmentation performance and generalization performance of the network.

This invention provides a method for cell nucleus segmentation in pathological images based on deep neural networks under weak supervision. The method includes: performing point annotation processing on sample images to generate a coarse supervision signal, including Venn diagram labels, cluster labels, and superpixel labels; using the Venn diagram boundary as a geometric prior, converting the superpixel labels into soft labels through an adaptive label smoothing strategy; constructing a segmentation network with an encoder-decoder structure, embedding a multi-domain edge module after each downsampling stage of the encoder to extract and enhance cell nucleus boundary features; embedding multi-faceted feature enhancement modules at the skip connections between the encoder and decoder to denoise, enhance, and structurally focus the features; and jointly training the network based on the supervision signal using a weighted multi-task loss function to obtain a trained segmentation model, which is then used to segment cell nucleus instances in the pathological image to be segmented. This invention achieves high-precision cell nucleus instance segmentation using only point annotations.

This invention discloses an infrared image analysis system for Czochralski silicon single crystal seed crystals, relating to the field of crystal growth process monitoring technology. The system includes an infrared image acquisition unit, an image enhancement module, a feature segmentation module, a graph construction and analysis module, and a quality assessment module. The infrared image acquisition unit acquires the original infrared image sequence; the image enhancement module uses a dual-stream Transformer network guided by illumination-independent features to suppress illumination interference and enhance structural features; the feature segmentation module achieves three-dimensional instance segmentation based on an improved Segment-Anything Model combined with multi-view fusion; the graph construction and analysis module maps instances to graph nodes, constructs an undirected sparse graph, and extracts topological geometric features; the quality assessment module determines the seed crystal state based on the feature vectors. This invention solves the problems of low infrared image contrast and difficult segmentation, achieving automated and high-precision quantitative assessment of the seed crystal fusion state.

The application discloses a kind of sparse point cloud and GCN fusion's citrus defect detection method and system, the point cloud of citrus data set is input into GCN_PointNet++ model in the application, the GCN_PointNet++ model combines GCN with PointNet++ network and carries out feature transmission by PaConv module;GCN_PointNet++ model is distilled by BIFPN, and the total distillation loss function is obtained by combining BIFPN distillation label to guide student network training, and the best point cloud segmentation result is obtained, that is, citrus defect segmentation point cloud model;Based on citrus defect segmentation point cloud model, three-dimensional reconstruction is carried out on citrus, three-dimensional point cloud mesh file is established, and the area of defects is calculated according to the surface formed by the mesh.The application realizes the accurate segmentation of citrus defect point cloud and the accurate quantification of citrus defect area.

The application discloses a coal mine image segmentation model and method based on a VMamba and multi-expert hybrid network and a construction method thereof. The coal mine image segmentation model is constructed. The image block encoding layer output of an encoder is taken as the input of the first encoding layer of a first VSS network, and the output of the first encoding layer of the first VSS network is taken as the input of the output end feature learning module. The input of the output end feature learning module is taken as the input of the first decoding layer of a decoder, and the output of the first decoding layer of the decoder is taken as the input of the second decoding layer of the decoder. The input of the second decoding layer of the decoder is taken as the input of a segmentation head, and the output of the segmentation head is taken as a segmented image. The application can significantly improve the segmentation precision and calculation efficiency of the coal image.

The application discloses an AI identification method, system and model training method for millimeter wave image overlapping targets based on a three-branch network, and solves the security target detection problem in millimeter wave images with low resolution, mutual interference or mutual overlap through collaborative work of three independent branch networks, namely, an analysis branch network, an aggregation branch network and a prediction boundary branch network. The analysis branch network saves the detailed information in a high-resolution feature map; the aggregation branch network realizes the feature position offset alignment of high and low resolution feature maps through a recursive fusion module; and the prediction boundary branch network extracts high-frequency features and enhances the perception and prediction of target edges. The multi-branch network structure effectively realizes the segmentation and identification of images through the fusion between branches. While improving the segmentation performance of the model, the model significantly reduces the calculation complexity and the detection error and missed detection probability in millimeter wave images, and finally realizes the accurate detection of overlapping objects.

The invention discloses a lumbar disc herniation interpretable diagnosis system based on cross-style consistency semi-supervised segmentation. The method comprises the following steps: firstly, carrying out style diversification processing on a lumbar axial MRI image through Bessel transformation to generate an image pair for consistency learning; the aggressive student model processes the stylized image and carries out cooperative training through cross pseudo supervision, and the robust student model provides a reliable prediction target through smooth parameter updating; the trained model can realize automatic segmentation of an intervertebral disc and a posterior spinous process area. Based on the segmentation result, the system adopts a geometric rule algorithm to quantify the prominence degree, and automatically generates a grading result and a visual report according to a clinical MSU grading standard. According to the method, the dependence on labeled data is reduced, the interpretability of the model is improved, and efficient and accurate lumbar disc herniation auxiliary diagnosis with clinical guiding significance can be realized.

The utility model discloses an arc extinguishing structure and a miniature circuit breaker, the arc extinguishing structure comprises an arc extinguishing cover and a separator, and the separator is connected to the rear end of the arc extinguishing cover; wherein the separator is provided with a first separating rib and a second separating rib, the first separating rib and the second separating rib are connected to divide the rear end of the arc extinguishing cover into four areas, and the four areas are respectively a left upper area, a left lower area, a right upper area and a right lower area; the arc extinguishing device has the advantage of being better in arc extinguishing effect.

This invention discloses a federated unified segmentation method and system with partial category labeling, relating to the field of image segmentation technology. The method includes: each client pre-training a local model using local sample data; the server distributing N pre-trained local models to each client; each client using its pre-trained local model to determine pseudo-labels for its local sample input data; training the client's corresponding pre-trained local model using the local sample data with determined pseudo-labels, obtaining the updated local model for the corresponding client; the server aggregating the N updated local models to obtain a current global model, iterating multiple times until the prediction accuracy of the current global model exceeds a set threshold, then using the current global model as a general global segmentation model; the general global segmentation model is used to segment the image to be segmented, outputting segmentation results for multiple categories. This invention improves the ability to segment multiple categories in an input image.

This invention belongs to the field of computer vision technology, specifically relating to a skin lesion image segmentation method based on a Transformer dual-branch model. The method constructs and trains a Transformer dual-branch model, inputting the image to be processed into the trained Transformer dual-branch model to obtain the segmentation result. The Transformer dual-branch model includes a main branch network, an auxiliary branch network, and an information aggregation module. This invention proposes a novel skin lesion image segmentation method that addresses the shortcomings of traditional deep learning methods in extracting global contextual information. It utilizes an efficient multi-scale visual Transformer as an encoder to extract more powerful and robust features. Simultaneously, it introduces low-level feature modules and high-level feature fusion modules to effectively improve the network's feature learning ability and segmentation performance.

The application discloses a point cloud segmentation method based on feature deviation values and an attention mechanism, relates to the technical field of three-dimensional point cloud classification, and utilizes the farthest sampling method to perform down-sampling on point clouds; a threshold radius is set with a sampling point as a center to extract feature points in the range by using a KNN algorithm; after the feature points are extracted, feature deviation values are calculated for each spherical neighborhood with the sampling center point, and then local features of the spherical neighborhood are obtained; residual multilayer perceptron shared weight parameters are utilized to fuse the local features by using attention pooling, and then global features of the point cloud are obtained; and the point cloud is segmented by utilizing the local features and the global features of the spliced point cloud. The application can effectively extract deep geometric features of the point cloud in a complex point cloud scene, fully excavate semantic information of the point cloud, avoid the problem of feature loss in a scene segmentation task, improve the segmentation capability of the model in a complex scene, effectively excavate deep information, and effectively realize scene segmentation of the point cloud.

The invention discloses an RGB-T semantic segmentation method and system based on multi-attention guidance and hierarchical fusion, and relates to the technical field of image processing. The system is composed of a double-flow encoder, a discriminative local texture perception unit, a semantic-driven cross-modal fusion unit, a semantic enhancement unit and a multi-scale layered refinement decoder, and efficient fusion and analysis of multi-modal features in a complex traffic scene are achieved. According to the discriminative local texture perception method, saliency features are learned through multi-attention guidance and a self-adaptive gating mechanism, accurate modeling of shallow texture information is focused, and the distinguishing ability of a region of interest and a target edge is improved; according to the semantic-driven cross-modal feature fusion method, efficient aggregation of global contexts is realized through high-level semantic guidance and cross-modal feature interaction, and feature complementarity is enhanced, so that the semantic-driven cross-modal feature fusion method has significant advantages in analysis of small targets, long-distance targets and boundary regions. The decoder adopts a progressive fusion mode, an additional edge detection module does not need to be added, and the overall segmentation precision is improved.

The application discloses an image semantic segmentation method, an electronic device and a computer readable storage medium, comprising: obtaining a target image and an image gradient map thereof; inputting the target image into a semantic segmentation network to obtain shallow feature maps and deep feature maps through an encoder; the semantic segmentation network comprises a cascaded encoder, a semantic-boundary double-branch decoder and an output layer; the feature fusion of multiple feature maps is performed through the semantic fusion branch of the decoder to obtain a semantic segmentation prediction result; the boundary information is obtained based on the image gradient map and the multiple feature maps through the boundary refinement branch of the decoder, and the feature fusion of the multiple feature maps is performed based on the boundary information to obtain a boundary prediction result; and the semantic segmentation prediction result and the boundary prediction result are fused through the output layer to obtain a semantic segmentation image of the target image. The application can further depict the object boundary in the semantic segmentation prediction result, thereby improving the accuracy and segmentation effect of semantic segmentation.

The application discloses a medical image tumor segmentation model based on a TransUNet framework and a construction method and a segmentation method thereof, the construction method comprising constructing a segmentation model based on the TransUNet, wherein an encoder of the segmentation model comprises CNN modules and a self-defined VSS module connected in sequence, the output ends of each convolution block of the CNN modules are input into a decoder of the segmentation model through a jump connection layer, the self-defined VSS module comprises a plurality of novel conversion layers stacked in series based on a Mamba module and a focus linear attention module, the self-defined VSS module performs multi-level feature interaction and global modeling on first encoding features output by the CNN module to obtain second encoding features; and the decoder performs feature fusion and up-sampling on the second encoding features and the output of the jump connection layer to obtain a segmentation image for a sample image. The application can significantly improve the segmentation robustness and precision of breast ultrasound images with low contrast, high noise and fuzzy boundaries.

The invention belongs to the field of financial big data analysis, and discloses a financial customer grouping method based on double-self-paced learning multi-view clustering under complex behavior data, and the method comprises the steps: constructing a multi-view anchor graph tensor of the complex financial behavior data, and comprehensively integrating the high-order consistent information of different financial behavior data; a self-paced learning strategy is introduced into tensor decomposition to gradually optimize a decomposition process, and double self-paced learning tensor decomposition is performed from two levels of samples and features, so that noise interference in Fourier domain transform is effectively reduced, and definition of a financial customer grouping structure and accuracy of a grouping result under complex behavior data are ensured. According to the method, tensor comprehensive analysis is carried out on multiple pieces of behavior data from the perspective of combining different views, a tensor decomposition process is optimized by using a self-paced learning strategy of a tensor level, a high-order interaction relationship among complex behavior data is effectively mined, and customer grouping in a complex financial scene is effectively carried out.

The application relates to a few-labeled remote sensing image semantic segmentation method and device based on visual text guidance, comprising the following steps: processing image data by using a visual feature encoder of a visual text model to obtain visual coding features; processing support image coding features and query image coding features by using a visual text prior decoupling model to obtain support image visual text priors and query image visual text priors; and obtaining mixed multi-level support image coding features by using a high-confidence visual feature model; obtaining visual relation prior values by using a multi-level prior calculation model; and decoding the priors by using a multi-level prior decoding network to obtain a segmentation result of the input image. In the application, visual text priors are introduced in remote sensing image small sample semantic segmentation, and the universality of the visual text model is used to solve the problem of large intra-class differences of remote sensing images in small sample semantic segmentation.

The invention discloses a real-time and off-line detection system and method for weld defects. The system comprises a planar electromagnetic chromatography sensor array, a multi-channel excitation and acquisition unit, a data processing and imaging unit, a prototype guide segmentation network unit and an information processing platform, the data processing and imaging unit is in communication connection with the prototype guide segmentation network unit and can input the preprocessed image into the prototype guide segmentation network unit, features are extracted through an encoder, spatial resolution is recovered step by step through a decoder, a prototype similarity graph is calculated through prototype branches, then the prototype similarity graph is fused with the features of the decoder, and a prototype image is obtained. And outputting the defect probability graph P and processing the defect probability graph P through an information processing platform to realize automatic judgment of weld defects. The method has the advantages that high robustness, high sensitivity and low omission ratio can be realized on the surface of a complex welding seam, and the technical problems that the detection sensitivity of superficial, fine and inclined defects is insufficient, real-time imaging and high-precision off-line reconstruction are difficult to consider, the omission ratio is difficult to reduce and the like are solved.

The application discloses a crystallization abnormality recognition method based on unsupervised learning, relates to the technical field of crystallization abnormality recognition, and comprises the following steps: setting a plurality of positions, acquiring crystallization video stream data, frame extraction and analysis to obtain multi-position crystallization frame images; each image is cut into a plurality of subblocks; an unsupervised learning algorithm is used to recognize the crystallization main body area of each subblock; feature extraction is performed on the crystallization main body area to obtain crystallization main body features; the similarity between the crystallization main body features of each subblock is compared, and the minimum similarity is taken as the overall similarity; the overall similarity is compared with an adaptive threshold value to determine whether there is crystallization abnormality. The application realizes all-around abnormality recognition by acquiring crystallization video stream data through multiple positions, and combines the unsupervised learning algorithm, feature extraction, similarity comparison and adaptive threshold value, so that the comprehensiveness and robustness of crystallization abnormality recognition are improved.

The application discloses an adaptive ship instance segmentation method based on feature decoupling, and relates to the technical field of image segmentation. In addition to utilizing a semantic main branch to extract multi-scale ship geometric feature maps, the adaptive ship instance segmentation network for ship instance segmentation also utilizes a weather perception auxiliary branch parallel to the semantic main branch to actively perceive the global environment in the image of the sea area to be segmented and generate an environment perception vector, thereby realizing double-path feature decoupling. A weather feature adaptive modulation module dynamically generates mapping parameters based on the environment perception vector, reconstructs and maps the ship geometric feature maps of the corresponding feature levels, and then performs instance segmentation, so that the feature distribution can be dynamically and adaptively adjusted according to the meteorological environment in the image of the sea area to be segmented, and the end-to-end inference efficiency, ship instance segmentation accuracy and multi-weather robustness are improved, and the method does not need to rely on complicated image defogging or rain removal preprocessing.

The application provides a fine brain region segmentation method, device, equipment and medium based on multiple levels and focus regions. The method comprises the following steps: acquiring a brain magnetic resonance image; performing multi-level segmentation on the brain region through a multi-level brain region segmentation network; according to a fine level brain region segmentation result, assigning different weights to each brain region of the fine level according to segmentation difficulty through a focus region segmentation network; and fusing the output results of the multi-level brain region segmentation network and the output results of the focus region segmentation network to obtain a final segmentation result of all brain regions. The application combines prediction on multiple levels, avoids training a separate network for each brain region, and can have good performance in terms of accuracy and efficiency. Moreover, the application proposes a loss based on focus regions, which helps to improve the segmentation capability of the network in difficult regions. Finally, a fusion module is used to fuse the predictions to obtain more accurate and robust brain region segmentation results.

The application relates to the technical field of image semantic segmentation, and discloses a multi-view joint semantic segmentation method for surround-view fisheye images, a multi-view joint segmentation model and a training process, which comprises the following steps: preparing a training data set, taking fisheye images under multiple views at the same moment as a group of surround-view images, the training data set comprising multiple groups of surround-view images and corresponding pixel-level semantic annotations; establishing a multi-view joint segmentation model based on real-time semantic segmentation, which comprises a feature extraction module, a distortion detail supplement module, a multi-view context fusion module and a decoder module; and training the multi-view joint segmentation model based on the training data set and cross-entropy loss. According to the application, multi-view distortion completion is first performed on high-resolution detail features in joint segmentation, and then multi-view position-related context fusion is performed in a deep layer, so that the correlation information between multi-view fisheye images is utilized, and high-precision semantic segmentation of the fisheye images is realized.