118 results about "Landmark point" patented technology

A highly reliable and high-precision navigation and positioning method and system for UAV under GPS-DENIED conditions includes: obtaining measurements of inertial sensor and image of binocular camera, extracting and tracking point features of the image, and obtaining altitude value; calculating position and attitude of camera and depth of landmark point based on the parallax and baseline of the binocular camera; fusing the measurement data of the inertial sensor and the binocular camera, and optimizing the data to obtain high-precision position and attitude data; obtaining the altitude value of the ranging radar and performing four-degree-of-freedom position and attitude graph optimization after adding the detected keyframes when repeated occurrence of the UAV at the same location is detected; encapsulating optimized position and attitude data to form a pseudo-GPS signal, inputting the pseudo-GPS signal to the UAV for positioning.

The invention provides a robot indoor navigation method and system, and the method comprises the steps: obtaining an RGB image frame sequence collected by a camera, and estimating a camera pose and a sparse three-dimensional landmark point of an image frame through a depth block visual odometer; based on the camera pose and the sparse three-dimensional landmark point, dense geometric enhancement is carried out on the image frame through a deep neural network of a transformer architecture to obtain dense geometric prior information; training a neural radiation field according to the initial camera pose, the dense geometric prior information and the image frame data, and performing joint optimization on neural radiation field parameters and the camera pose based on luminosity consistency, geometric consistency and camera pose constraints; based on image frames collected in real time, novel view synthesis is carried out through the optimized nerve radiation field, and robot navigation is carried out based on a novel view. According to the scheme, accurate estimation of the camera pose and high-quality reconstruction of the three-dimensional scene can be realized, the computing resource overhead is low, and the real-time performance can be guaranteed.

The invention provides a three-dimensional scene reconstruction method and apparatus, and an XR device, and relates to the technical field of image processing, the three-dimensional scene reconstruction method is applied to the XR device, and comprises the steps of receiving a scene image acquired by the XR device; performing feature extraction and feature matching on the scene image to obtain a matching point pair; converting the matching point pairs into three-dimensional landmark points under a global coordinate system; and performing gridding reconstruction according to the three-dimensional landmark points under the global coordinate system to obtain a three-dimensional grid model. According to the method, real-time reconstruction of the indoor scene three-dimensional model can be realized on the premise of not depending on a special depth camera.

A system and method of landmark detection using deep neural network with multi-frequency self-attention is provided. The system includes an encoder network that receives an image of an object of interest as an input and generates multi-frequency feature maps as output. The system further includes an attention layer that receives the generated multi-frequency feature maps and refines the generated multi-frequency feature maps based on correlations or associations between the received multi-frequency feature maps. The system further includes a decoder network that receives the refined multi-frequency feature maps as a second input from the attention layer and generates a landmark detection result based on the second input. The landmark detection result includes a heatmap image of the object of interest and the heatmap image indicates locations of landmark points on the object of interest in the image.

The invention discloses an oral cavity modeling method, system and device and a nonvolatile storage medium. The method comprises the steps that multiple frames of first scanning data, obtained on the basis of a first condition, of the oral cavity are obtained, and the multiple frames of first scanning data comprise oral cavity morphology three-dimensional data and first mark point three-dimensional data; second scanning data, obtained on the basis of a second condition, of the oral cavity are obtained, and the second scanning data comprise three-dimensional data of a second mark point; and obtaining a target oral cavity model based on the first scanning data and the second scanning data. The technical problems of inaccurate intraoral data and large oral cavity modeling error caused by gingiva deformation and soft tissue jittering during scanning of an intraoral scanner are solved.

The invention relates to the crossing field of biomedicine and image processing technology, and discloses a focused ultrasound mouse brain tissue targeted positioning method and positioning system based on magnetic resonance guidance and application, the positioning system comprises a data layer, a model layer, an acquisition layer, an image layer and an algorithm layer, the data layer is used for processing a total data set J, the model layer is used for calculating and obtaining a brain tissue mark point prediction position, the acquisition layer is used for obtaining data for calculating a mark point prediction model M0, and the image layer is used for carrying out space correction on a current brain tissue magnetic resonance image A and aligning a corrected brain tissue magnetic resonance image A0 with an image template B. And the algorithm layer provides an algorithm basis for correcting the difference between brain tissue individuals and carrying out local coordinate adjustment on the predicted position of the mark point. According to the invention, magnetic resonance guidance is used for focused ultrasound mediation related operation, and target brain tissues can be accurately targeted and positioned.

The invention relates to the technical field of orthodontics and medical image processing, in particular to an artificial intelligence-based automatic positioning method and system for a maxillary bone expansion device. The positioning method comprises the following steps: taking a cone beam computed tomography image and intraoral scanning data of a patient; segmenting a key anatomical structure based on the scanned image, wherein the key anatomical structure comprises an incisor foramen and a transverse palate slit; automatically identifying a plurality of cranio-maxillofacial mark points based on the scanned image, and establishing a maxilla three-dimensional coordinate system; based on the three-dimensional coordinate system, the intraoral scanning data and the key anatomical structure, positioning adjustment is carried out on the arch expander to determine the final positioning posture, and positioning adjustment comprises the steps that the safe distance between the arch expander and the surface of the palate mucosa is screened, and contact with the key anatomical structure is avoided; and outputting the three-dimensional parameters of the final positioning posture. According to the invention, the automatic, intelligent and safe placement of the arch expanding device is realized, so that the design efficiency is improved, the operation difficulty is reduced, and the clinical safety and the arch expanding effect are guaranteed.

The application provides a method and system for spatial calibration of a radar device. The method comprises: obtaining an image and radar measurement data at the same time, the radar measurement data comprising radar coordinates of a plurality of landmark points in a geographic region; obtaining geographic coordinates of the landmark points recorded in the image according to pixel coordinates of the landmark points recorded in the image; determining at least two landmark point matching pairs, each landmark point matching pair comprising a geographic coordinate of a landmark point and a radar coordinate of the landmark point; and obtaining a calibration relationship between the radar measurement data and the geographic region according to the at least two landmark point matching pairs. The method can improve the accuracy and efficiency of spatial calibration.

The invention discloses a mark point layout method and device for rotation monitoring and a medium. The method comprises the steps of obtaining a curved surface normal vector and a curved surface center point of a curved surface structure to be monitored; acquiring the optimal position of the binocular camera; obtaining a threshold value of a sight angle of the binocular camera according to the eccentricity of the elliptical distortion of the circular mark point; obtaining the relationship between the visibility of points on each position of the curved surface and the rotation angle of the curved surface structure; obtaining the relationship between the elliptical distortion and the rotation angle of the circular mark point; obtaining an image available frame of each point location under the sight line of the current rotation angle; and constructing mark point selection logic to obtain an optimal layout position of the mark points. The positioning method of the binocular camera and the positioning method of the position of the mark point are adopted to optimize the layout of the mark point, the elliptical distortion of the mark point is reduced, the precision of rotation monitoring is improved, the layout of the mark point can be optimized, the measurement blind area is avoided, the number of the mark point is reduced, the time for pasting and removing the mark point in the measurement operation is shortened, and the measurement efficiency is improved.

The invention is a system and method of introducing an automated detection system for manual and semi-automated visual inspection processes. The system employs a facial tracking camera to monitor an operator's gaze direction and level of attention during the inspection process. One or more images of a subject are received, and facial landmarks / associated confidence values are output. The landmark points and any associated confidence values, and / or associated information, may then be input to another set of one or more models that may output any facial analysis quantity or quantities. The system can analyze the data and trigger an alert if it detects an attention lapse, allowing for immediate intervention.

This application belongs to the field of computer vision and optical measurement technology, specifically disclosing a point cloud stitching method and apparatus based on monocular pose prior marker matching. The method includes: calculating the pose of multiple markers extracted from left and right camera images, determining the normal vector and three-dimensional coordinates of each marker's circular cross-section in its respective camera coordinate system; transforming the three-dimensional coordinates and corresponding normal vectors of each marker in either the left or right camera image to the other camera coordinate system for marker point registration between the left and right camera images; the circular cross-section is a circular cross-section in three-dimensional space with the origin of the camera coordinate system as its vertex, passing through the elliptical curve where the marker is located; and stitching the three-dimensional point cloud of markers based on the marker point registration results of the left and right camera images. This application can significantly improve the accuracy of binocular stereo marker matching and the system's anti-interference capability in complex industrial scenarios.

A computing device obtains an image of a user and detects a facial region of the user within the image. The computing device detects facial landmark points within the facial region and extracts facial features based on the detected facial landmark points. The computing device calculates traits of the user based on the extracted facial features and calculates one or more personality types of the user based on the traits of the user.

The application provides a triangular facet data management method, device and XR equipment, relates to the technical field of data processing, and the method comprises the following steps: performing spatial grid division on initial three-dimensional landmark points to construct a grid map; performing surface reconstruction on the three-dimensional landmark points in the grid map to obtain triangular facet data of each grid; obtaining vertex indexes of the triangular facets according to the triangular facet data; and calculating hash indexes of the triangular facets according to the vertex indexes of the triangular facets, so as to map the triangular facets to a triangular facet index list of the corresponding grid. The application can solve the problem that the existing triangular facet data management method is difficult to realize fast and accurate query of triangular facet data.

The invention relates to an iterative optimization-based binocular camera-IMU external parameter on-line calibration method, which does not need to manually set a checkerboard calibration plate, and adopts an on-line iterative calibration method to directly iteratively calibrate rotation external parameters and translation external parameters of a left camera and an IMU in a binocular camera until convergence. Monocular vision inertia beam adjustment optimization is carried out based on the rotation external parameter and the translation external parameter of the left camera, 3D point coordinates of the landmark point are obtained, and the rotation external parameter and the translation external parameter between the left camera and the right camera are calibrated based on the 3D point coordinates of the landmark point of the left camera and the 2D point coordinates in the right camera. Compared with the prior art, the method has the advantages of saving time and material cost and the like.

A method performed by an electronic device for estimating a marking point of a body part of a subject by the electronic device is provided. The method includes generating, by the electronic device, an initial coarse estimate of the marker point of the body part using a lightweight deep neural network, determining, by the electronic device, an occlusion region of the body part based on the generated initial coarse estimate of the marker point using a segmentation mask, estimating, by the electronic device, an occlusion probability for the flag point in the at least one occlusion region and the generated initial coarse estimate, determining, by the electronic device, a correction factor for application to the generated initial coarse estimate as a measure of the estimated occlusion probability, and selecting, by the electronic device, a predefined number of neural networks by applying the determined correction factor for processing the at least one occlusion region and the generated initial coarse estimate to generate a final estimate of the flag point.

The application relates to a landmark point matching method and device, computer equipment and a storage medium. The method comprises the following steps: obtaining a scanning point cloud obtained by shooting a calibration object under a current scanning visual angle; the calibration object comprises a landmark point; matching the calibration object point cloud with the scanning point cloud to obtain a first matching point pair; the first matching point pair comprises a calibration object matching point and a matching scanning matching point; projecting the scanning matching point to an image coordinate system of a camera to obtain an image mapping point; obtaining a calibration object image obtained by shooting the calibration object by the camera; matching the image mapping point with the landmark point on the calibration object image to obtain a second matching point pair; the second matching point pair comprises a scanning matching point and a matching landmark point on the calibration object image; for the camera, determining a target matching point pair of the landmark point on the calibration object image and the landmark point on the calibration object based on the first matching point pair and the second matching point pair. The method can obtain more accurate target landmark point pairs and optimize unstable factors of the equipment.

Systems and methods of landmark detection using deep neural networks with multi-frequency self-attention are provided. The system includes an encoder network that receives an image of an object of interest as an input and generates a multi-frequency feature map as an output. The system also includes an attention layer that receives the generated multi-frequency feature map and refines the generated multi-frequency feature map based on a correlation or association between the received multi-frequency feature maps. The system further includes a decoder network that receives the refined multi-frequency feature map from the attention layer as a second input and generates a landmark detection result based on the second input. The landmark detection result includes a heat map image of the object of interest, and the heat map image indicates locations of landmark points on the object of interest in the image.

A method of generating a correction plan for correcting a deformed bone includes inputting to a computer system a first image of the deformed bone in a first plane and inputting to the computer system a second image of the deformed bone in a second plane. Image processing techniques are employed to identify a plurality of anatomical landmarks of the deformed bone in the first image. The first image of the deformed bone is displayed on a display device. A graphical of the deformed bone is autonomously generated and graphically overlaid on the first image of the deformed bone on the display device, the graphical template including a plurality of lines, each line connected at each end to a landmark point corresponding to one of the anatomical landmarks. A model of the deformed bone may be autonomously generated based on the graphical template.

The invention relates to an SLAM back-end optimization method based on sparse matrix factorization and a hardware accelerator architecture, and the method comprises the steps: constructing a Hessian matrix through employing a binary sparse coding mechanism according to the actual observation coordinates of feature points on an image plane and the predicted coordinates of road sign points projected to a camera coordinate system; performing Schel elimination processing on the Hessian matrix based on binary sparse coding to obtain a linear equation set only containing a camera pose variable; and carrying out extended Kalman filtering on a linear equation set only containing a camera pose variable to realize incremental updating of a state pose and a road sign point coordinate. According to the method, the overall delay of SLAM rear-end optimization can be remarkably reduced, and the real-time requirement in a high-speed moving scene is met.

The invention requests to protect an unsupervised anatomical mark point accurate positioning method based on double-layer alignment and a super-resolution head, and aims to solve the problem of domain migration caused by differences of different clinical center imaging devices. Comprising the following main steps: S1, constructing an input layer alignment module, generating a target style image by adopting adaptive instance normalization, and meanwhile, retaining an original spatial anatomical structure; s2, constructing an output layer alignment module, adopting a mean value teacher framework, and guiding student model learning through a high-quality pseudo tag and a self-adaptive mark point shielding strategy; and S3, introducing a lightweight super-resolution head, fusing the multi-scale features extracted by the backbone network, and generating a high-resolution heat map by adopting pixel rearrangement. Compared with an existing method, the method has the following advantages that (1) the cross-domain adaptive capacity is effectively improved through input-output double-layer alignment; (2) the super-resolution head can directly decode the coordinates without tedious post-processing, so that the quantization error and the calculation complexity are reduced;

A method of generating a correction plan for correcting a deformed bone includes inputting to a computer system a first image of the deformed bone in a first plane and inputting to the computer system a second image of the deformed bone in a second plane. Image processing techniques are employed to identify a plurality of anatomical landmarks of the deformed bone in the first image. The first image of the deformed bone is displayed on a display device. A graphical of the deformed bone is autonomously generated and graphically overlaid on the first image of the deformed bone on the display device, the graphical template including a plurality of lines, each line connected at each end to a landmark point corresponding to one of the anatomical landmarks. A model of the deformed bone may be autonomously generated based on the graphical template.

The embodiment of the invention discloses a data processing method and device, computer equipment and a storage medium. The method is applied to a scanning scene of a target scanning object, the target scanning object is provided with a target or a first mark point installed in a mouth, the interior of the mouth of the target comprises the first mark point, and the method comprises the following steps: obtaining mouth scanning data of the target scanning object, the oral scanning data comprises intraoral scanning data and first mark point scanning data of the first mark point; acquiring surface scanning data of the target scanning object, wherein the surface scanning data comprises the first mark point scanning data; and aligning the mouth scanning data with the surface scanning data according to the first mark point scanning data to obtain target data. According to the embodiment of the invention, data alignment is carried out through the public mark points (the first mark point scanning data) in the mouth scanning data and the surface scanning data, so that accumulated errors can be reduced, and the splicing precision between the mouth scanning data and the surface scanning data is improved.

The invention relates to the technical field of medical data processing, and discloses a three-dimensional lower jawbone mark point determination method, system and device based on a thermal point cloud network and a medium, and the method comprises the steps: calculating the Euclidean distance between a three-dimensional lower jawbone point cloud data midpoint and a three-dimensional lower jawbone mark point, obtaining the Gaussian distribution, and determining the three-dimensional lower jawbone mark point. The relevancy between the points in the three-dimensional lower jawbone point cloud data and the three-dimensional lower jawbone mark points is represented, and better guidance is provided for training of a lower jawbone mark point prediction model. The mandible mark point prediction model performs local feature extraction of three-dimensional mandible point cloud data through a local feature extraction module based on a PointNet + + architecture by using a thermodynamic point cloud network, performs global feature extraction in combination with a global feature extraction module based on the PointNet architecture, and outputs a three-dimensional thermodynamic diagram and a coordinate deviation value through feature fusion and decoding; the three-dimensional lower jawbone mark point can be accurately predicted, and reliable data support is provided for medical research.

A muscle movement information-containing intraoral scanning data accurate acquisition method comprises the following steps: S1, collecting a single-frame texture map including an intraoral photo or a dental scanning image, marking an attached gum boundary by using a dental professional, creating a ground truth segmentation map, and performing creation and training of an attached gum area automatic segmentation model after performing variant on U-Net; s2, generating edge line data by utilizing Poi ntNet + + based on gypsum model scanning and artificially described edge lines, and creating and training an automatic edge line generation model; the model aims at automatically generating accurate edge lines for the intraoral scanning data obtained in the step S4 and guiding manufacturing of the movable restoration; s3, dynamically recording three-dimensional form information of intraoral muscles of the user in a motion state; s4, the obtained dynamic data and conventional static oral scanning data are guided by mark points to be elastically registered, and an intraoral data sequence with the complete form is obtained.According to the method, the implementation scheme that oral scanning is used for replacing impression taking can be achieved.

This application relates to a marker point matching method, splicing method, apparatus, device, and storage medium. The method includes: obtaining a current frame marker point set and a buffer frame marker point set; the current frame marker point set includes current frame marker points; the buffer frame marker points in the buffer frame marker point set are selected from global marker points; obtaining a first distance matrix corresponding to the current frame marker points and a second distance matrix corresponding to the buffer frame marker points; based on the first and second distance matrices, obtaining a pair of matching points with the same name for the current frame marker points and the buffer frame marker points; the matching point pair includes the same-name points of the current frame marker points; updating the buffer frame marker points based on the same-name points of the current frame marker points to obtain updated buffer frame marker points; the updated buffer frame marker points are used for matching with marker points in the next frame. This method can shorten the marker point matching time, thereby improving real-time performance.

The application provides a digital oral model mark point recognition method and device and electronic equipment, and the method comprises the following steps: step 1, inputting dental arch model image data, segmenting the dental crown of the dental arch model, and recognizing the tooth position number of the dental crown, wherein the dental arch model comprises a gum and dental crowns of different tooth positions. Step 2, based on automatic recognition of the dental crown mark point of a single dental crown, for each segmented dental crown, the multi-view graph network is used to automatically recognize the dental model mark point, and the mark point on each dental crown is obtained. The method of the application does not predict the mark point on the three-dimensional point cloud, but maps the three-dimensional data into two-dimensional projection image data, and calculates by using a two-dimensional network, so that the calculation amount and the dependence on the computing power can be reduced compared with a three-dimensional network.

A landmark registration registers landmark points of navigation guidance information for performing a navigational guidance of a surgery procedure. The registration operation comprises a landmark point identification procedure and a landmark matching procedure. In particular, the present case is directed to automatically identifying three or more landmark points in an anatomy image by using a model according to shapes of a bone ridge depicted in the anatomy image. A graphical user interface enables interactively confirming locations of the three or more landmark points in the anatomy image. Given the identified three or more landmark points, the present case further describes interactively matching respective locations of the respective landmark points in the anatomy image with physical locations of a probe tip in anatomy according to deviation information of the probe tip. The automated identification and the interactive matching of landmark points enable generating navigational surgical guidance information with accuracy.

The method includes accessing a plurality of reference landmark points associated with a plurality of features in a reference image, accessing a reference measurement point corresponding to an asset in the reference image, receiving a target image including the features and the asset, identifying a plurality of target landmark points of the target image corresponding to the features, generating a mapping from the reference landmark points to the target landmark points, determining, using the mapping, a target measurement point corresponding to the asset in the target image, and extracting, from the target measurement point of the target image, a temperature measurement of the asset.

The invention relates to the technical field of image positioning, and discloses a 3DGS visual repositioning method and system based on intrinsic image decomposition, a terminal and a storage medium, and the method comprises the steps: obtaining multi-view scene image data, inputting the multi-view scene image data into a three-dimensional Gaussian sputtering model for training, and obtaining an intrinsic 3DGS map model; obtaining a query image to be positioned and a landmark point set, inputting the query image into the internal 3DGS map model for decomposition to obtain a target albedo map, and performing feature matching according to the target albedo and the landmark point set to obtain an initial pose; and rendering the initial pose based on an internal 3DGS map model to obtain a dense candidate albedo map and a depth map, and adjusting the initial pose according to the dense candidate albedo map, the depth map and the target albedo map to obtain a positioning camera pose. According to the invention, visual repositioning is carried out on the query image through the internal 3DGS map model, and accurate visual positioning is realized.

The invention belongs to the technical field of three-dimensional measurement, and particularly discloses a high-robustness mark point matching method and device, and the method comprises the steps: obtaining a distance matrix under adjacent visual angles based on a mark point set S and a mark point set T under the adjacent visual angles; searching potential matching point pairs between S and T based on a preset screening condition and the distance matrix under the adjacent view angles; respectively screening X mark points closest to the center of the mark point from S and T, and putting the X mark points into two candidate point sets; based on a preset screening condition and the distance matrixes corresponding to the two candidate point sets, candidate point pairs between the two candidate point sets are searched, and a pair of reference points is screened out from the searched candidate point pairs; and based on the found potential matching point pair and the found pair of reference points, through matching point pair search based on congruent triangles, screening out two pairs of reference points from the found potential matching point pair. According to the invention, the robustness of three-dimensional data splicing can be effectively improved.