Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

605results about How to "Accurate segmentation" patented technology

Auricle anaphora segmentation method and system

The invention provides an auricle anaphora segmentation method and system, and the method comprises the steps: inputting a to-be-segmented human ear image and text description into an auricle anaphora segmentation model which comprises a text coding module, a text-guided visual coding module and visual decoding module, and an angle transformation module; the text coding module obtains a text feature # imgabs0 #; a visual coding module realizes fusion of text features and image features through a structure organized into four stages, a visual encoder in each stage generates visual features # imgabs1 #, a cross-modal sensing module aligns # imgabs2 # and # imgabs3 # to obtain multi-modal features # imgabs4 # and # imgabs5 #, each element in the multi-modal features # imgabs4 # and # imgabs5 # is weighted by an attention gating module to obtain weighted multi-modal features # imgabs6 #, and the weighted multi-modal features # imgabs6 # and # imgabs5 # are weighted by an attention gating module. Adding the elements and # imgabs7 # to generate an enhanced visual feature # imgabs8 #, inputting # imgabs9 # into a visual decoding module to gradually recover the spatial resolution of the image, and further fusing the text and the visual feature to output a multi-scale feature; the angle transformation module carries out angle transformation on the multi-scale features and outputs a segmentation mask of a text description related region. According to the invention, the auricle can be subjected to anaphora segmentation.
Owner:UNIV OF SCI & TECH BEIJING

Power battery laser welding defect detection method based on three-dimensional vision

The invention discloses a power battery laser welding defect detection method based on three-dimensional vision, and the method comprises the steps: carrying out the noise reduction of obtained laser welding point cloud data through radius filtering, and carrying out the rough segmentation of the laser welding point cloud data after noise reduction through a random sampling consensus algorithm, accurately segmenting the laser welding point cloud data subjected to coarse segmentation by using a curvature threshold self-adaptive algorithm to obtain welding point cloud data; performing defect segmentation on the welding seam point cloud data through an improved region growing algorithm to obtain welding seam defect point cloud data; and on the basis of the welding seam defect point cloud data, through a preset welding seam defect classification and size calculation method, obtaining the type and size of the welding seam defect. Accurate segmentation and size measurement of small-size defects in welding seams can be realized.
Owner:HUBEI UNIV OF ARTS & SCI +1

AI-based medical image automatic segmentation and labeling system

The invention relates to the technical field of medical treatment, in particular to an AI-based medical image automatic segmentation and labeling system, which comprises an image data acquisition module, a data preprocessing module, a feature extraction module, an automatic segmentation module and an intelligent labeling module, the data preprocessing module improves image quality through noise suppression, artifact removal and modal alignment, the feature extraction module extracts a feature map of a potential lesion area, and the automatic segmentation module combines a full convolutional network, a bidirectional long-short-term memory network and an attention mechanism to realize accurate segmentation of an organ area and the lesion area. The intelligent labeling module is used for labeling anatomical names, lesion types and pathological features of the segmentation results by using a medical knowledge graph and a semantic embedding technology; the method has the advantages of being high in segmentation precision, comprehensive in marking and high in expansibility, and the automation and intelligence level of medical image analysis can be improved.
Owner:HUNAN INSTITUTE OF ENGINEERING

Image segmentation method and device combining feature difference recognition and detail enhancement

The invention provides an image segmentation method and device combining feature difference recognition and detail enhancement, and the method comprises the following steps: obtaining a to-be-segmented image to construct a data set, and carrying out the marking and preprocessing of the data set; constructing an image segmentation model, wherein the image segmentation model comprises an encoder network, a jump connection network, a decoder network and an output network; the encoder network uses a pre-trained CNN (Convolutional Neural Network) and Transform hybrid network to carry out feature extraction; the jump connection network comprises a channel space interaction module for fusing encoder features and decoder features; the decoder network comprises a convolution-channel attention module for extracting key features and a convolution-up-sampling module for up-sampling; the output network comprises a multi-level feature fusion module for fusing the multi-scale features of the decoder; training the image segmentation model to obtain a trained image segmentation model; and segmenting the image data by adopting the trained image segmentation model.
Owner:SICHUAN UNIV

Few-sample defect detection method based on prototype prompt fine-tuning visual basic model SAM

The invention belongs to the technical field of image recognition, and discloses a few-sample defect detection method based on a prototype prompt fine-tuning visual basic model SAM. Prototype prior information about an image to be segmented is extracted from a small number of example image-mask pairs, so that prior guidance of human beings is replaced; for defect data, a representative foreground prompt point, a high-confidence background prompt point (sparse prompt) and a foreground Mask (dense prompt) are extracted from prototype prior to serve as prototype prompt, and a visual basic model SAM is guided to achieve more accurate segmentation. A visual basic model SAM is finely adjusted by using prototype prompt and a small amount of defect data, a prompt encoder and an image encoder of the SAM are frozen, an LoRA bypass is injected into the image encoder, and the SAM is enabled to better adapt to defect data distribution by updating parameters of the LoRA bypass and a lightweight mask decoder.
Owner:NORTHEASTERN UNIV CHINA

Pyramid vision Transform-based polyp image segmentation method

The invention relates to a polyp image segmentation method based on pyramid vision Transform, and belongs to the technical field of digital medical treatment. The method comprises the following steps: capturing global context information through a pyramid vision Transform module, and obtaining feature maps with different resolutions from a plurality of extraction stages; detail information is extracted from the shallow feature map of the first stage through a cyclic cross attention module, and shallow features are obtained; a multi-scale feature fusion sensing module is used for carrying out fusion processing on the feature maps extracted in multiple stages, deep features are extracted, and the deep features and shallow features are fused and then used for image segmentation; and designing weighted binary cross entropy loss and weighted IoU loss to train the above process, and segmenting the polyp image after training is completed. According to the invention, through the pyramid vision Transform module, the cyclic cross attention module and the multi-scale feature fusion perception module, the global information learning ability, the detail information capturing ability and the segmentation performance are significantly improved.
Owner:CHONGQING UNIV OF POSTS & TELECOMM

Image segmentation-based vertebral body three-dimensional reconstruction method, bone detection method and system

The invention relates to the technical field of image processing, and provides a vertebral body three-dimensional reconstruction method based on image segmentation, and a bone detection method and system.The method comprises the steps that firstly, an SAD-UNetR image segmentation model capable of achieving accurate segmentation of an image target is constructed through improvement of a Swinin-UNetR network, three-dimensional modeling is conducted on the basis of an accurate segmentation image, and a three-dimensional reconstruction model is obtained; an accurate three-dimensional model of a vertebra structure can be obtained; based on an accurate segmentation result, a bone mineral density prediction regression model is provided at the same time, bone substances are identified based on multi-instance learning and an attention mechanism, a bone substance classification result and bone mineral density data can be provided, intelligent assistance is provided for diagnosis of doctors, radiation risks, economic consumption and time consumption caused by special bone mineral density measurement are avoided, and the accuracy of bone mineral density measurement is improved. Measurement means of bone detection are expanded, and the method has a wide application prospect.
Owner:SHANDONG NORMAL UNIV

Perception method based on lightweight real-time instance segmentation and semantic mapping

The invention belongs to the technical field of computer vision, provides a real-time semantic perception system running on equipment, performs real-time instance segmentation and semantic mapping on a working environment, and provides a perception method based on lightweight real-time instance segmentation and semantic mapping. In the first part, firstly, a color image containing color and texture information is used for object classification and bounding box regression; cutting the original depth image into a depth image with an accurate geometric edge, and inputting the depth image into the truncation network to generate a mask of the instance; a second part: (1) calculating the attitude and size of a target object in the view of the camera through principal component analysis; (2) obtaining the category of the object in the view obtained by segmentation; (3) creating a point cloud of the depth image projection of the target object in a map; and then cross-frame tracking and fusion are carried out by using data association to realize the purpose of sensing. The method is mainly applied to automatic sensing occasions.
Owner:TIANJIN UNIV

Mobile phone screen glass defect detection method based on improved YOLOv8

The invention discloses a mobile phone screen glass defect detection method based on improved YOLOv8, and relates to the technical field of mobile phone screen glass cover plate quality detection. The method comprises the following steps: simulating an industrial production environment, collecting mobile phone screen glass images, and constructing a data set; improvement and optimization are carried out on the basis of the YOLOv8 target detection network model, a mobile phone screen glass defect detection classification model is constructed, and parameters are trained and optimized; constructing a semantic segmentation model of the YOLOv8 network, segmenting the defects of the mobile phone screen glass image to be identified, and realizing detection and identification of different defect category edges; and fitting the actual damage size of the mobile phone screen glass defect range. The mobile phone screen glass defect detection system is suitable for accurate detection of mobile phone screen glass defect characteristics, can measure the defect range, and provides intelligent and efficient services for the mobile phone screen glass quality inspection process.
Owner:ANHUI AGRICULTURAL UNIVERSITY

Pavement detection method based on artificial intelligence and road inspection robot

The invention discloses a road surface detection method based on artificial intelligence and a road inspection robot, and the method employs a self-designed RCDIT network model, proposes an implicit expression method and a multi-head cross attention mechanism in the network model, guarantees the expression integrity and continuity of image blocks, and improves the detection precision. Complex interaction of image blocks and position information can be effectively processed, and the integration capability and information relevance of the model to multi-modal data are enhanced. In addition, the introduced position feedforward network further improves the abstract ability of feature representation, and is helpful for more effectively processing complex features and improving the expressive force of the model. And the RCDIT network model is combined with the road detection robot, so that the crack defects of the road surface and the road can be accurately and quickly identified and segmented, and the reliability and the accuracy of detection are greatly improved. The road detection robot can work in a 24 / 7 all-weather mode and is not limited by time and environment, manpower resources are effectively saved, and continuous monitoring on the road surface can be achieved.
Owner:EAST CHINA JIAOTONG UNIVERSITY

Fruit stem pose recognition method applied to pomelo picking equipment

The invention discloses a fruit stem pose recognition method applied to pomelo picking equipment, and the method comprises the following steps: carrying out the preprocessing of image data of a to-be-recognized scene, and obtaining the preprocessed image data; the image data comprises an RGB image and a depth image; inputting the preprocessed image data into the optimized instance segmentation model to obtain a depth map of a single pomelo fruit; on the basis of the depth map of the single pomelo fruit, an elliptical fruit posture feature region is recognized, and an elliptical inclination angle of the fruit posture feature region is obtained in combination with a Hough ellipse detection algorithm; and calculating the position and posture of the pomelo stem according to the ellipse inclination angle of the fruit posture feature region. According to the method, through integration of multiple advanced technologies, the whole process optimization from image data preprocessing to fruit stem pose accurate identification is realized, the identification precision and the working efficiency of pomelo fruit picking automation equipment are improved, and the false picking rate and the fruit damage rate are reduced.
Owner:SOUTH CHINA AGRICULTURAL UNIVERSITY

Single tree crown segmentation deep learning method based on multi-modal data

The invention discloses a single tree crown segmentation deep learning method based on multi-modal data, and belongs to the technical field of artificial intelligence in forestry scientific research. Based on a high-resolution image GF-2, a laser radar Li DAR point cloud feature and ultrahigh-resolution RGB multi-modal data, a vectorization tree crown post-processing strategy is fused, a deep learning model, namely a multi-modal tree crown division fusion network (MTCDFNet) is proposed, and a deep learning model is constructed by integrating a forest structure, spectrum and texture features of multi-modal remote sensing data. Accurate segmentation of the crowns in the multi-canopy and multi-tree-species subtropical forest is realized.
Owner:RES INST OF FOREST RESOURCE INFORMATION TECHN CHINESE ACADEMY OF FORESTRY

Semantic segmentation method, system and equipment for aerial image and medium

The invention aims to provide an aerial image semantic segmentation method, system and equipment and a medium, and mainly solves the problem of low semantic segmentation precision caused by the fact that modal difference between a visible light image and a thermal infrared image is not considered in the prior art. A convolutional network model based on an attention mechanism is designed for semantic segmentation of an RGB-T image, and the model adopts a double-flow convolutional neural network encoder to perform feature extraction and complementary perception fusion on the RGB-T image obtained by aerial photography of an unmanned aerial vehicle. And through the channel attention module and the space attention module, the weights of different modal data in different channels are captured, and the spatial resolution and the positioning precision are improved, so that the RGB-T image semantic segmentation model has the characteristics of high image data feature utilization rate and good semantic segmentation precision under different illumination conditions.
Owner:ZHEJIANG UNIV OF TECH

Peanut segmentation method based on improved YOLOv8n-seg

The invention discloses a peanut segmentation method based on improved YOLOv8n-seg. The method comprises the following steps: collecting a peanut image, carrying out data enhancement, and marking to generate a peanut data set; dividing the peanut data set into a training set and a verification set according to a set proportion; an improved YOLOv8n-seg model is constructed, a deformable convolution DCNv3 module, a C2fDCNv3 module, a scale sequence feature fusion SSFF module, a triple feature encoder TFE module and an EMA attention module are introduced, and an improved YOLOv8n-seg-DSE model is obtained; using the training set and the verification set to train and verify the YOLOv8n-seg-DSE model, and obtaining a peanut segmentation model; and obtaining a to-be-segmented peanut image and inputting the to-be-segmented peanut image into the peanut segmentation model to obtain a peanut segmentation result According to the invention, the YOLOv8n-seg model is improved, the extraction of peanut features is enhanced, the peanut segmentation precision is improved, the size of the model is reduced, and the model is easier to deploy at a mobile terminal.
Owner:CHINA JILIANG UNIV

Sock flaw detection method and system based on machine vision

The invention relates to the technical field of textile defect detection, in particular to a sock defect detection method and system based on machine vision, and the method comprises the steps: firstly, respectively segmenting a to-be-detected sock image and a corresponding flawless sock image into a plurality of regions through constructing a multi-stage region segmentation algorithm, and obtaining a to-be-detected region set and a flawless region set; then, respectively extracting a fabric structure shape of each region in the to-be-detected region set and the flawless region set, and obtaining a to-be-detected fabric structure shape set and a standard fabric structure shape model of each region; then, detecting defect spots by calculating the shape difference between the structure shape of each to-be-detected fabric in the region and the corresponding standard fabric structure shape model; and finally, integrating flaw spot detection results of all areas to obtain a sock flaw detection result. According to the invention, the precision and reliability of flaw detection can be effectively improved under the background of various sock fabric structures.
Owner:ZHUJI ZHISHEN SOCKS IND CO LTD

Laser welding quality evaluation method

The invention relates to the technical field of image processing, in particular to a laser welding quality evaluation method, which comprises the following steps of: acquiring a grey-scale map of a welding image, dynamically adjusting the size of an initial region of each pixel point to determine a local region of each pixel point, forming a feature descriptor of each pixel point by using feature values of multiple dimensions of the local region, and determining the welding quality of the welding image according to the feature descriptor. Calculating the similarity of all the pixel points based on the feature descriptors of all the pixel points, performing clustering based on the similarity, and calculating the weight of each category after clustering; and according to the weight of each category and the average feature value of the feature descriptors of the pixel points contained in the category in each dimension, adjusting the initial adaptive segmentation threshold of each pixel point, performing adaptive threshold segmentation, and evaluating the welding quality based on the result of the adaptive threshold segmentation. And the accuracy of adaptive threshold segmentation and the accuracy of a welding quality evaluation result are improved.
Owner:WUXI CHAOQIANGWEIYE TECH CO LTD

Water body remote sensing recognition method based on hyperspectral data simulation and U-net

The invention discloses a water body remote sensing recognition method based on hyperspectral data simulation and U-net. The water body remote sensing recognition method comprises the following steps: A, data collection and preprocessing; b, selecting and manufacturing a sample area; c, data simulation; d, derived parameters are calculated; e, manufacturing a training sample and a verification sample; f, training a U-net network model; g, multispectral water body detection; and H, precision evaluation. According to the method, a high-quality training sample is obtained through a hyperspectral remote sensing image, multispectral data of different sensors are simulated, the problem that the reusability of water body classification samples among different multispectral satellites and at different geographic distances and time is not sufficient is solved, the multispectral data Landsat8 and Sentinel2 are taken as examples, and the reusability of the water body classification samples is improved. It is proved that by means of the method, easily-confused ground objects such as building shadows and cloud shadows in the water body recognition process can be effectively removed, and on the basis of accurately extracting large-area rivers and lakes, the recognition precision of small lakes and ponds is improved.
Owner:BEIJING XINGDI VISION TECH CO LTD +1

Lupus nephritis image segmentation method, lupus nephritis image segmentation device, medium and equipment

The invention discloses a lupus nephritis image segmentation method and device, a medium and equipment, and relates to the technical field of image segmentation. In order to solve the technical problem of inaccurate segmentation of lupus nephritis images due to local optimum, the invention provides a method. The method comprises the following steps: obtaining a to-be-segmented lupus nephritis gray image and a corresponding two-dimensional histogram; searching a group of segmentation thresholds on the two-dimensional histogram by using an improved rime algorithm; the improved rime algorithm comprises the steps that a foraging-storage optimization stage parallel to an original rime algorithm optimization stage is added, and the original rime algorithm optimization stage and the foraging-storage optimization stage are executed randomly; and performing threshold segmentation on the to-be-segmented lupus nephritis grayscale image based on the segmentation threshold. According to the method, different optimization stages are executed through random lines, so that the optimization process is prevented from falling into a fixed search mode, the algorithm is prevented from being stopped near a certain local optimal solution, and the opportunity of finding a threshold which enables lupus nephritis image segmentation to be more accurate is increased.
Owner:BIG DATA & INFORMATION TECH RES INST OF WENZHOU UNIV

Microcirculation blood vessel automatic segmentation and quantification method based on enhanced feature fusion

The invention discloses a microcirculation blood vessel automatic segmentation and quantification method based on enhanced feature fusion. The method comprises the following steps: step 1, data preparation and preprocessing; 2, designing an enhanced feature fusion network model; step 3, automatically segmenting microcirculation blood vessels; and 4, quantifying the diameter of the blood vessel based on the segmentation result. According to the enhanced feature fusion network model provided by the invention, accurate segmentation of microvessels is realized through fusion and enhancement of multi-scale features. The model shows good generalization ability on diversified clinical image data through effective data enhancement, and can adapt to image data with different illumination conditions, contrast ratios and resolutions. Therefore, the model has high practicability and reliability in practical application. According to the method, the diameter of the blood vessel is quantified through the segmentation result, the diameter of the microvessel can be accurately measured, and powerful support is provided for evaluation of the clinical microcirculation function.
Owner:HARBIN INST OF TECH AT WEIHAI +1

Industrial image defect detection method based on attraction and rejection contrast learning

The invention provides an industrial image defect detection method based on attraction and rejection contrast learning, and relates to the technical field of industrial defect detection. The method comprises the following steps: acquiring normal image data from an existing image data set; performing data enhancement on the obtained normal image data to obtain a synthesized abnormal image; the synthesized abnormal images are respectively sent to a teacher network and a student network of an industrial image defect detection model for feature extraction, and multi-level feature fusion is respectively carried out; gaussian noise is added to the fused features of the student network, and then the features are sent to a UNet network for de-noising processing; and finally, attraction and rejection processing is carried out on data of the two networks, the target that the difference between normal pixels becomes small and the difference between abnormal pixels becomes large is achieved, model parameters are kept unchanged in the test stage, and the defect position is positioned by calculating the output feature difference of the two networks. The method has a better defect positioning effect and a higher reasoning speed, and can meet the requirements of industrial production.
Owner:BOHAI UNIV

Short video annotation method based on video questions and answers

The invention relates to the technical field of video questions and answers, in particular to a short video labeling method based on video questions and answers, which comprises the following steps of: S1, collecting short video materials from a plurality of short video platforms according to different video types, and setting each type of video extraction frame frequency by combining the duration of each video according to different types of videos; s2, for each video frame, extracting high-dimensional feature representation of a visual object and a scene text, and extracting feature vector sequences of different modals by using a multi-modal transformer; and S3, through a pre-trained deep learning model, identifying an object in the video frame, and further identifying scene change in the video in combination with the scene detection model. The invention provides a short video annotation method based on video questions and answers, and the method achieves the precise segmentation of movie and television contents through the steps of multi-modal data collection, preprocessing, semantic analysis, multi-modal fusion, bridge segment segmentation and the like. The system can accurately identify semantic boundaries in complex film and television contents, and has a wide application prospect.
Owner:BEIJING GUANJIANG SPACE TECH CO LTD

Polyp segmentation method based on edge guidance and deep supervision PVT

The invention discloses a polyp segmentation method based on edge guidance and deep supervision PVT. The method aims at a colonoscope image with a polyp, and outputs the position and the shape of the polyp by using a constructed polyp segmentation network. The polyp segmentation network uses PVT-v2 as a backbone network for feature extraction, and further comprises a context fusion decoding branch, a boundary extraction branch, a boundary sensing module and a cascade feature fusion branch. And the context fusion decoding branch serves as an image decoder and receives a semantic feature map output by the PVT-v2 backbone network. And the boundary extraction branch enriches the content of the detail feature map by using a refining module to obtain an edge prediction result. And the boundary sensing module fuses the refined features and the output feature map of the context fusion decoding to obtain a boundary sensing prediction result. And the cascaded feature fusion branch fuses fusion features under different scales layer by layer according to a sequence from deep to shallow, and outputs a final polyp segmentation prediction result P5.
Owner:HANGZHOU DIANZI UNIV

Image threshold segmentation method, system and device for defect detection of display screen and storage medium

The invention discloses an image threshold segmentation method, system and device for defect detection of a display screen, and a storage medium, and aims to realize efficient, accurate and robust defect detection, facilitate improvement of an image segmentation effect, accurately segment a defect region of an image, and facilitate analysis of a defect image. The method comprises the steps of collecting an initial image displayed under a display screen through a camera; performing down-sampling processing on the initial image to obtain a down-sampled image which is a part of the initial image; carrying out expansion filling on the edge of the downsampling image to obtain an expanded image; performing preprocessing and filtering processing on the extended image to obtain a filtered image; and dynamically generating a selection window according to a window selection strategy, and selecting a target image from the filtered image through the selection window.
Owner:SHENZHEN SEICHITECH TECHN CO LTD

Remote sensing image ship segmentation method based on U-KAN network model

The invention discloses a remote sensing image ship segmentation method based on a U-KAN network model. The method comprises the following steps: constructing the U-KAN network model; the U-KAN network model is a hybrid network architecture fusing a U-Net network and a KAN network; wherein the U-Net network comprises an encoder and a decoder which are in jump connection; a plurality of KAN layers are nested in the KAN network; u-KAN network model training is carried out in combination with the remote sensing image data set, and an optimized U-KAN network model is obtained; and inputting a remote sensing image to be processed into the optimized U-KAN network model to obtain a corresponding ship segmentation result. The method not only can effectively capture ship targets with different scales and complexities in the remote sensing image, but also can realize effective fusion between shallow detail features and deep semantic features, thereby overcoming the limitation in processing complex backgrounds and diversified ship forms in the prior art.
Owner:耕宇牧星(北京)空间科技有限公司

Intelligent data analysis system and method based on aluminum oxide production

The invention relates to the technical field of intelligent analysis of aluminum oxide, in particular to an intelligent data analysis system and method based on aluminum oxide production. Acquiring ball mill parameters, raw material parameters and a spherical alumina impurity threshold, and determining a standard threshold according to the ball mill parameters and the raw material parameters; sampling and detecting the processed spherical aluminum oxide to obtain a purity and morphology image set of the spherical aluminum oxide; constructing a morphology image segmentation model to segment the morphology image set to obtain a first image set and a second image set; constructing a first image recognition model to recognize the first image set to obtain a spherical alumina parameter set, and calculating to obtain a first coefficient; constructing a second image recognition model to recognize the second image set to obtain an impurity parameter set, and calculating to obtain a second coefficient; and calculating the quality coefficient according to the first coefficient, the second coefficient, the spherical alumina impurity threshold and the spherical alumina purity. The comprehensive evaluation of the spherical aluminum oxide is realized through the quality coefficient of the spherical aluminum oxide.
Owner:JIANGSU JINGJING NEW MATERIALS

MRI (Magnetic Resonance Imaging) brain tumor segmentation method based on multi-scale attention mechanism and secondary segmentation strategy

The invention discloses an MRI (Magnetic Resonance Imaging) brain tumor segmentation method based on a multi-scale attention mechanism and a secondary segmentation strategy, and belongs to the field of semantic segmentation. The method comprises the following steps: firstly, acquiring MRI brain tumor data and preprocessing the MRI brain tumor data; secondly, designing a 3D U-Net model, and combining a multi-scale attention mechanism and depth separable convolution to extract features; then, secondary segmentation guided by a cascade strategy is introduced; and finally, predicting by using the optimal model, storing and verifying a result, and calculating and comparing evaluation indexes. A depth separable convolution module is used for replacing a traditional convolution module, so that the parameter and calculation cost is reduced, and the segmentation precision is kept. In addition, a multi-scale attention fusion module is introduced to the first two layers of the encoder, the ability of capturing shallow information is enhanced, and important features are further extracted. A cascade attention mechanism is introduced into a decoder, and the segmentation capability of a small target area is improved. And training is carried out through a BraTS2020 data set, and the result shows that the optimized network significantly improves the MRI brain tumor segmentation accuracy.
Owner:CHONGQING UNIV OF TECH

Gesture data processing method in motion, electronic equipment and storage medium

The invention discloses a gesture data processing method in motion, electronic equipment and a storage medium, and belongs to the technical field of user interaction. In order to achieve accurate data processing of gestures in a motion state, the method comprises the steps that an intelligent data glove is used, and three-axis accelerometer data in the motion process and three-axis gyroscope data in the motion process are collected; comparing and analyzing the acquired data of the three-axis accelerometer in the movement process and the acquired data of the three-axis gyroscope in the movement process, and extracting gesture recognition data used in the movement process; constructing an arm state segmentation algorithm module based on arm motion state detection, and inputting gesture recognition data used in motion into the arm state segmentation algorithm module based on arm motion state detection to obtain filtered dynamic gesture data; and inputting the filtered dynamic gesture data into a dynamic gesture segmentation algorithm module based on line segment fitting model comparison to perform gesture segmentation, and then performing data noise reduction to obtain processed gesture data in motion.
Owner:HARBIN INST OF TECH

Nerve block anesthesia ultrasonic guided data processing method

The invention relates to a data processing method for nerve block anesthesia ultrasonic guidance. The method comprises the following steps: firstly, preprocessing ultrasonic image data, removing image noise by adopting a noise reduction algorithm based on deep learning, and enhancing key structures including visibility of nerves and blood vessels by using a local contrast enhancement algorithm; secondly, performing feature extraction by using deep learning, identifying and segmenting nerves, blood vessels and muscle tissues in the image, and simultaneously integrating data of different ultrasonic modes to extract tissue features; then, the processing flow is optimized through a recursive learning algorithm and a self-adaptive algorithm, the recursive learning algorithm optimizes the accuracy of feature extraction and image analysis according to the previous injection effect and the ultrasonic image, and the self-adaptive algorithm adjusts processing parameters according to the body types and tissue characteristics of different patients; and finally, constructing a complete three-dimensional tissue model from the continuous two-dimensional ultrasonic image sequence by utilizing a volume reconstruction algorithm, and controlling the positioning and real-time navigation of a needle head in combination with a real-time position sensing technology.
Owner:南昌大学第一附属医院

Engineering structure crack detection and measurement method based on machine vision

The invention belongs to the technical field of machine vision and deep learning, and discloses an engineering structure crack detection and measurement method based on machine vision, and the method comprises the steps: collecting original image data of an engineering structure, discretizing the original image data into image blocks, carrying out the preprocessing, building a crack classification model through a Resnet-50 neural network, and carrying out the training, the method comprises the following steps of: firstly, constructing an image block, identifying a crack image in the image block, then constructing a SegFormer segmentation model based on a Transformer framework to segment the crack image to obtain a crack mask graph, evaluating the performance of the two models, then analyzing the two models by adopting an orthogonal skeleton line method, obtaining a contour line and a skeleton line of a crack, and calculating the length of the crack. According to the method, the ResNet-50 image classification model and the SegFormer segmentation model are combined to realize high-precision crack identification and segmentation of the building structure crack, the contour line and the skeleton line of the crack are obtained and the length of the crack is measured by means of the orthogonal skeleton line method, and the crack of an engineering structure is well detected and subjected to risk assessment.
Owner:SANYA SCI & EDUCATION INNOVATION PARK WUHAN UNIV OF TECH +1

Intelligent segmentation method and system for space steel structure nodes

The invention discloses an intelligent segmentation method and system for space steel structure nodes, and belongs to the technical field of building construction. The method comprises the following steps: collecting point cloud data of a steel structure, and preprocessing the point cloud data; clustering the point clouds to form a plurality of super voxels, and calculating concavity and convexity between adjacent super voxels; abstracting boundaries between adjacent super voxels into point clouds with convex-concave information, endowing the concave points and the convex points with weight values, and calculating an optimal segmentation plane; according to the method, the super voxels are segmented according to the optimal segmentation plane, and the segmentation result is optimized, so that the nodes are accurately segmented from the whole steel structure model, and the nodes are conveniently monitored by using a finite element model. The method can achieve the intelligent and rapid recognition and segmentation of the complex space steel structure nodes, remarkably improves the recognition and segmentation efficiency of the nodes, can guarantee the high consistency of segmentation results, improves the precision of the segmentation results, and can guarantee the comprehensiveness and accuracy of damage detection.
Owner:SHANGHAI CONSTRUCTION FOURTH CONSTRUCTION GROUP CO LTD