70results about How to "Good registration effect" patented technology

The invention provides a precise registration method of multilook point cloud, comprising the following steps: respectively selecting one piece of point cloud overlapping approximately from two piecesof global point cloud to be registered to serve as the target point cloud and the reference point cloud; utilizing a principle direction bonding method to realize the preregistration of the target point cloud and the reference point cloud; utilizing the principle direction test method to judge and realize the consistency of the preregistration principle directions of the target point cloud and the reference point cloud; respectively calculating the curvature of each point in the target point cloud and the reference point cloud; respectively obtaining characteristic matching point symetries P0and Q0 according to curvature similarity; using the iterative closest point matching algorithm to realize the precise registration of the target point cloud and the reference point cloud by utilizingthe characteristic matching point symetries P0 and Q0; and completing the registration of the two pieces of global point cloud. The method is characterized by high computation speed and high registration precision, thus being capable of realizing good registration effect.

A system and method for image registration includes tracking (508) a scanner probe in a position along a skin surface of a patient. Image planes corresponding to the position are acquired (510). A three-dimensional volume of a region of interest is reconstructed (512) from the image planes. A search of an image volume is initialized (514) to determine candidate images to register the image volume with the three-dimensional volume by employing pose information of the scanner probe during image plane acquisition, and physical constraints of a pose of the scanner probe. The image volume is registered (522) with the three-dimensional volume.

The invention discloses a cross-modal medical image registration method. The method comprises the following steps: providing a training set comprising a floating image of a first modal and a referenceimage of a second modal; inputting the floating image into an image conversion network, and converting the floating image into a conversion image of a second modal; inputting the floating image and the reference image into a cross-modal flow sub-network to output a first deformation field; inputting the converted image and the reference image into the single-mode flow sub-network to output a second deformation field; inputting the first deformation field and the second deformation field into a deformation field fusion network to output a final deformation field; inputting the floating image and the final deformation field into a space transformation network to obtain a floating image after distortion transformation of the final deformation field; obtaining a first total loss function according to the transformed floating image and the reference image, and performing supervised training on the network by taking minimization of the first total loss function as a target; and inputting ato-be-registered image into the trained network to obtain a registered image. The cross-modal medical image registration method can greatly improve the cross-modal medical image registration effect.

The embodiment of the invention provides a medical image processing method, device and equipment. The method comprises the steps: obtaining a first medical image and a second medical image of a specified target; registering the second medical image to the first medical image based on a first spatial transformation model, and obtaining a first registering image; registering the first registering image to the first medical image based on a second spatial transformation model through an unsupervised deep learning method, obtaining a second registering image, and enabling the fine structures in the specified target to be aligned through the registering of the unsupervised deep learning method, thereby reducing the artifacts, and improving the precision of the registering result. Especially for a tissue comprising many fine structures, the method can obtain a better registering result, and solves a problem to certain degree that the precision of a registering result is lower for the tissue comprising many fine structures in the prior art.

The invention belongs to the technical field of three-dimensional reconstruction, and discloses a local reference point-based quick point cloud registration method. The method comprises the steps of performing down-sampling on input original points cloud to obtain sparse point clouds corresponding to the original point clouds; obtaining local included angle features by taking points in the sparsepoint clouds as reference points; establishing a kd-tree by the local included angle features of a view angle; searching for nearest neighbor features of the local included angle features of the otherview angle; taking a k1 pair with a minimum Euclidean distance error in the nearest neighbor features as a seed matching pair; searching for other matching pairs under the reference from neighborhoods; calculating out an adjacent view angle transformation matrix by using rigid body transformation to finish rough registration of the point clouds; and optimizing a rough registration result by usinga partially overlapped ICP registration algorithm. According to the method, initial matching is established by utilizing quick included angle features, and the matching neighborhoods are used for propagation, so that the registration precision is ensured while the point cloud registration speed is increased.

The invention discloses a femoral or tibial registration method and device based on total knee replacement surgery. The method comprises the steps of acquiring data point pairs, wherein the data pointpairs include a CT data reconstructed femoral or tibial digital three-dimensional model and selected feature points corresponding to bone points; adopting a point pair registration algorithm for performing rough registration on the data point pairs to obtain rough registration results; acquiring data point cloud pairs, wherein the data point cloud pairs include the CT data reconstructed femoral or tibial digital three-dimensional model and slidably collected point clouds corresponding to bones; performing precise registration according to the rough registration results and the data point clouds. According to the method and the device, anatomical feature points which are clear for a doctor are adopted for carrying out a rough registration algorithm, and dense three-dimensional point cloudstereoscopic data is further utilized to perform a fine registration algorithm, so that the registration accuracy rate is greatly increased. The first registration result based on SVD-ICP is used as the initial position of the next ICP registration, and the point cloud data is collected again to achieve a better registration effect from a more satisfactory initial position.

The invention provides a multi-modal image registration method based on a synthetic ultrasonic image. The method comprises the following steps: according to a magnetic resonance image and a real ultrasonic image, constructing a generative adversarial network comprising a generator and a discriminator, generating a simulated synthetic ultrasonic image, registering the simulated synthetic ultrasonicimage and a real ultrasonic image to obtain registration parameters, and applying the registration parameters to a magnetic resonance image to complete registration and fusion of the final magnetic resonance image and the final ultrasonic image. According to the method, the ultrasonic image can be synthesized from the magnetic resonance image in real time, and the requirement of a real-time imageguide operation is met; the synthesized simulated ultrasonic image is closer to a real ultrasonic image, the image quality is higher, and important detail information is better stored; when the magnetic resonance image contains the tumor, the simulated ultrasonic image can still be accurately synthesized; the final registration technology does not need a complex registration algorithm, and a goodregistration effect can be achieved only through a traditional simple registration algorithm.

The invention discloses a gradient distribution-based non-rigid medical image registration method. The steps are as follows: (S1) a normalized mutual information similarity measure is constructed; (S2) a gradient distribution distance is constructed; (S3) a non-rigid registration model is created; (S4) the registration model is solved. The gradient distribution-based non-rigid medical image registration method not only utilizes the gray distribution characteristics of images to be registered, but also takes the gradient distribution information of the medical images into consideration, thus effectively fulfilling the non-rigid registration of the medical images. Moreover, preprocessing, such as feature extraction and splitting, does not need to be carried out on the images to be registered, but the joint probability distribution of gray values of two images is utilized to calculate normalized mutual information between the images to be registered, the gradient distribution-based non-rigid medical image registration method also has a good registration effect on medical images with great grayscale difference, and adaptivity and robustness are high.

The invention brings forward a novel image registering method. The method comprises the following steps: establishing gray-scale-based mutual information registering adaptation value function; establishing and initializing a group, wherein four dimensions of each individual in the group respectively represents horizontal translation, vertical translation, a rotation angle and a zoom coefficient of a floating image; according to the mutual information registering adaptation value function, calculating a fitness value of each individual, and calculating an optimal position of the whole group; by use of an iteration mechanism of a differential evolution algorithm, updating a position vector of each individual, and updating the optimal position of the whole group; determining whether conditions for executing an alternative strategy is satisfied, and if so, executing the corresponding alternative strategy; and repeatedly executing the aforementioned steps until maximum iteration frequency Tmax of the differential evolution algorithm is satisfied. The novel image registering method has the advantages of good registering stability and high precision, greatly improves the performance of an image registering algorithm and lays a reliable foundation for subsequent image processing work.

The invention discloses an improved artificial bee colony algorithm-based image registration method. The method includes the following steps that: (1) parameters to be optimized and a fitness function are determining, and a first generation artificial bee colony is initialized, so that an initial food source can be generated; (2) employed bees find adjacent food sources of the current food source, and select a food source between the current food source and the adjacent food sources through adopting a greedy method; (3) the employed bees share food source information with observation bees, and the observation bees calculate the fitness values of all the food sources; (4) the observation bees select food sources with small fitness values through utilizing a roulette algorithm, and perform secondary search for adjacent food sources through adopting difference strategies; (5) step (2) to step (5) are repeated until a maximum number of times of iteration is obtained; and (6) an optimal solution is outputted. With the improved artificial bee colony algorithm-based image registration method adopted, the process of image registration can be accelerated, and the accuracy of the image registration can be improved.

The invention discloses a registering and compression method of a three-dimensional grid sequence. By means of the method, the redundancy of the grid sequence on inter-frame topological information can be effectively removed, the grid sequence compression efficiency can be greatly improved, so that storage space is reduced. The method includes the steps that firstly, the three-dimensional grid sequence is registered based on inter-frame coding of a video; secondly, vertexes and patch data of all I frames are subjected to lossless compression through open source software 7-Zip, and vertex data of P frames and vertex data of B frames are compressed by compressing vertex residual data between adjacent frames, wherein patch data of the P frames and patch data of the B frames can be shared by the I frames; thirdly, a compressed file in the second step is decompressed through the open source software 7-Zip, and then reconstruction is carried out.

The invention provides a method for multispectral image registration and a device for the multispectral image registration. The method comprises determining characteristic points in a reference image A and a to-be-registered multispectral image B, calculating a main direction of the reference image A and a main direction of the to-be-registered multispectral image B, enabling the reference image A to rotate in the main direction of the reference image A, enabling the to-be-registered multispectral image B to rotate in the main direction of the to-be-registered multispectral image B, extracting characteristic points after rotation, performing filtering on the characteristic points after the rotation, calculating an edge region EOH descriptor of each characteristic point, matching the characteristic point, which has the minimum Euclidean distance of the EOH descriptor from a to-be-matched characteristic point in the to-be-registered multispectral image B, in the reference image A with the to-be-matched characteristic point, and performing registration on the reference image A and the to-be-registered multispectral image B according to the matched characteristic points in the reference image A and the to-be-registered multispectral image B. The method for the multispectral image registration and the device for the multispectral image registration which are provided in the invention improve a registration effect of a multispectral image in the rotation condition, and improve registration precision of the multispectral image.

The invention discloses a tooth occlusion analysis system based on point cloud spatial characteristics. The system comprises a data acquisition unit, a preprocessing unit, an overlapping region positioning unit, a point cloud registration unit and an occlusion analysis unit; the data acquisition unit is used for acquiring tooth jaw point cloud data and tooth point cloud data of the same case in different periods; the preprocessing unit is used for extracting and obtaining auxiliary point cloud data based on the tooth jaw point cloud data and the tooth point cloud data; the overlapping region positioning unit determines an overlapping region according to the tooth jaw point cloud data, the tooth point cloud data and the auxiliary point cloud data; the point cloud registration unit completes registration of dental jaw point cloud data of the same case in different periods based on the overlapped area; and the occlusion analysis unit performs tooth occlusion analysis based on the registered tooth jaw point cloud data. According to the method, the influence of low overlapping rate and noise on the matching effect can be reduced to the maximum extent, and the registration precision is greatly improved.

The invention provides a point cloud registration method and device, which combine point cloud semantic segmentation and rasterization, consider the attribute characteristics of a point cloud, providericher characteristic information except for height and reflectivity in point cloud data, and can adopt different registration schemes for different attribute characteristics, thereby improving the matching accuracy. On one hand, the attribute features serve as the basis of registration, and for some point cloud data with few strong feature scenes, the matching accuracy can also be improved; andon the other hand, the semantic segmentation can also remove noise of vehicles, pedestrians and the like which interferes with matching, so that the matching accuracy is further improved. For a scenewith a large collected data error (greater than 5m), the point cloud registration effect is particularly obvious by adopting the scheme of the embodiment of the invention. The invention further provides a server and a computer readable medium.

The invention discloses a ground laser scanning three-dimensional point cloud data registration method. The method comprises the following steps: establishing a point cloud registration model; and scanning original three-dimensional point cloud data under the relative coordinate system of the excavation surface of the earth material field, and calculating the theta around the Z axis between the relative coordinate system xoy of the registration model and the world coordinate system XOY, so that point cloud registration can be carried out. According to the point cloud data registration method, the registration efficiency is extremely high, a large amount of time can be saved in practical application, and time cost consumption can be reduced for engineering practical application.

The present application relates to an image registration method, comprising: selecting a registration source image {p,} and a registration target image {qj}; applying a random perturbation to {pi} in accordance with a preset random perturbation control parameter σ so that it is deformed to obtain {pi′}, and obtaining a set of closest points, i.e., {qj′}, on {qj} corresponding to points on {pi′}; performing an iterative operation on {pi} and {qj′} in accordance with a preset initial coordinate transformation H0 to obtain a coordinate transformation {Hl}, 0<l<L, between {pi′} and {qj′}, and calculating an average distance {El} corresponding to the coordinate transformation {Hl} in accordance with the coordinate transformation {Hl}; judging magnitudes of the average distance {El} and a preset ideal average distance Ex, and terminating the registration when the average distance {El} is smaller than or equal to the preset ideal average distance.

The invention is applicable to the field of computers, and provides an MR image and CT image registration method, which comprises the steps of obtaining an MR image and a CT image to be registered; generating a simulated CT image according to the MR image; determining a current displacement field according to the simulated CT image and a CT image to be registered; processing the current displacement field and the previous displacement field based on a differential homoembryonic algorithm to generate a superposition state displacement field; performing regularization processing on the superposition state displacement field to generate a smooth deformation field; generating a registered to-be-confirmed image according to the smooth deformation field and the simulated CT image; and until a preset condition is met, determining that the registered to-be-confirmed image is a registered image, otherwise, returning to the step of determining the current displacement field according to the analog CT image and the to-be-registered CT image to perform iteration processing. According to the registration method provided by the embodiment of the invention, the MR image is processed into the analog CT image, and then iterative optimization registration is carried out, so that the registration method also has a good registration effect on the image with relatively large deformation.

The invention discloses a medical image registering method by means of an improved NNDR (nearest neighbor distance ratio) strategy, and belongs to the medical image registering field. The medical image registering method by means of the improved NNDR strategy provides an improved matching method INNDR based on NNDR, and increases the point correct matching situation of next nearest neighbor. The medical image registering method by means of the improved NNDR strategy can significantly improve the number of correctly matched characteristic point so as to obtain a better registering effect. The experimental result shows that compared with the original method and other criterions, the medical image registering method by means of the improved NNDR strategy has significant advantages. The medical image registering method by means of the improved NNDR strategy can find more matched pairs and can improve the accuracy of the finally matching. The medical image registering method by means of theimproved NNDR strategy uses the reference image to test the method, and verifies the matching result by means of quantitative evaluation. The experimental result shows that the medical image registering method by means of the improved NNDR strategy has relatively higher accuracy and effectiveness.

The invention provides an image registration method and system based on Gaussian field constraint and manifold regularization. The method is used for establishing correct matching by removing the wrong matching in the initial matching point pair, includes establishing a non-rigid geometric transformation model corresponding to the geometric transformation between images to be matched, establishingan objective function based on Gaussian field constraint and manifold regularization, and solving a parameter matrix by a deterministic annealing method; calculating the non-rigid geometrical transformation model by using the model parameters, and judging the correctness of the initial matching point pair according to the threshold value; and finally, establishing the mapping relationship betweenthe two images by using the correct set of matching points, and realizing the image registration. The invention models the non-rigid transformation between the images to be registered, and establishes a robust estimator, which greatly reduces the error rate of the matching, and still maintains good robustness even in the case of a large number of error matches in the preliminary matching.

The invention provides a precise registration method of multilook point cloud, comprising the following steps: respectively selecting one piece of point cloud overlapping approximately from two pieces of global point cloud to be registered to serve as the target point cloud and the reference point cloud; utilizing a principle direction bonding method to realize the preregistration of the target point cloud and the reference point cloud; utilizing the principle direction test method to judge and realize the consistency of the preregistration principle directions of the target point cloud and the reference point cloud; respectively calculating the curvature of each point in the target point cloud and the reference point cloud; respectively obtaining characteristic matching point symetries P0 and Q0 according to curvature similarity; using the iterative closest point matching algorithm to realize the precise registration of the target point cloud and the reference point cloud by utilizing the characteristic matching point symetries P0 and Q0; and completing the registration of the two pieces of global point cloud. The method is characterized by high computation speed and high registration precision, thus being capable of realizing good registration effect.

The invention discloses a registering and compression method of a three-dimensional grid sequence. By means of the method, the redundancy of the grid sequence on inter-frame topological information can be effectively removed, the grid sequence compression efficiency can be greatly improved, so that storage space is reduced. The method includes the steps that firstly, the three-dimensional grid sequence is registered based on inter-frame coding of a video; secondly, vertexes and patch data of all I frames are subjected to lossless compression through open source software 7-Zip, and vertex data of P frames and vertex data of B frames are compressed by compressing vertex residual data between adjacent frames, wherein patch data of the P frames and patch data of the B frames can be shared by the I frames; thirdly, a compressed file in the second step is decompressed through the open source software 7-Zip, and then reconstruction is carried out.

The invention discloses an image registration method and device, electronic equipment and a computer readable storage medium. The invention relates to the image registration method. The method comprises the following steps: carrying out global registration on a reference image block in a reference frame and a to-be-registered image block in a to-be-registered frame, and determining a global registration result of the reference image block; performing block matching on the reference image block in the reference frame and the to-be-registered image block in the to-be-registered frame, and determining a block matching result of the reference image block; and determining a target image registration result of the reference frame according to the global registration result and the block matching result. According to the invention, registration effects of the global registration result and the block matching result can be comprehensively considered, and the target image registration result of the reference frame is determined; the accuracy of image registration of the reference frame and the to-be-registered frame is improved, and a noise reduction effect of an image obtained by fusing the reference frame and the to-be-registered frame is further improved; and meanwhile, the registration effects of the global registration result and the block matching result are comprehensively considered, so that the noise reduction effects at boundaries between different regions on the image obtained by fusion are small in difference.

The invention discloses a three-dimensional multi-modal image deformable registration method for a rigid target. A point cloud is constructed from a multi-modal three-dimensional image, a local feature coordinate system is constructed by using normal and curvature information, and an iterative nearest neighbor algorithm is corrected and error points are eliminated, so that the registration robustness is improved. Partition registration is carried out based on the structure information of the point cloud, the difference among the regions is met on the basis of local rigid registration, deformable rigid registration is achieved, and the overall registration precision is improved. An existing iterative nearest neighbor registration method is insufficient in utilization of local features of point clouds and has large errors. Therefore, multiple features of the point clouds are fully utilized to construct the local feature reference coordinate system to improve the registration process, and point clouds are subjected to partition registration in combination with the structural features of the point clouds, so that a better registration effect of registering the point clouds and the reference point clouds is achieved.

The invention provides a high-precision screen printing screen plate scanning image registration method. According to the invention, based on own features of screen plate imaging, selecting a registration feature point suitable for the screen plate image feature; and then carrying out geometric transformation on the screen image based on a planar two-dimensional affine transformation method, so that the screen image has a good registration effect with the template image after transformation, the problem of large calculation amount is solved, overfitting is avoided, and the real-time effectiveness of registration is ensured.

The invention discloses a three-dimensional image iterative registration method, which comprises the following steps: (1) acquiring a reference image and a target image, calculating the central coordinates of each vertebral section of the reference image and the target image, and obtaining an initial transformation relationship between the reference image and the target image; (2) generating a plurality of reference cross-section images in the reference image, and intercepting the target image according to the step (1) to obtain a corresponding target cross-section image; (3) performing optimal similarity matching on a certain target cross-section image and the corresponding reference cross-section image to obtain the transformation quantity, and calculating to obtain the pose of the corresponding target cross-section coordinate system; (4) judging whether the pose increments of the plurality of target section coordinate systems are all smaller than a set threshold, and if so, calculating a final transformation relation between the reference image coordinate system and the target image coordinate system; otherwise, updating the transformation quantity and the pose of the corresponding target section coordinate system, and repeating the step. According to the invention, accurate registration between the two vertebral section images can be rapidly completed, the efficiency is high, the adaptability is high, and a good registration effect can be achieved for poor image data.

The invention discloses a point cloud registration strategy in flatness measurement of a laser contourgraph, and solves the problem of rapid and accurate estimation of a point cloud pose of an object to be measured in the flatness measurement of the laser contourgraph. The method comprises the following steps: firstly, extracting a boundary of a point cloud through an algorithm to obtain a more effective key point SHOT feature descriptor, and eliminating invalid key points through key point neighborhood evaluation to obtain a preliminary matching point pair; secondly, clustering the feature matching point pairs in combination with geometric consistency of the feature matching point pairs so as to realize coarse registration of the point clouds; and finally, completing fine registration of the point clouds by using an ICP algorithm based on the boundary point clouds. In comparison experiments with classical point cloud registration algorithms such as SHOT + ICP, 4PCS + ICP, NDT and the like, the speed and the precision of the method are improved to a certain extent.

The invention provides a safe fingerprint verification method and a system, and a safe fingerprint verification system is constructed by using an image acquisition unit, a verification characteristic extraction unit, an assisted characteristic extraction unit, a template encryption unit, a template storage unit, a registration parameter computing unit, a template decryption unit and the like. The method and the system provided by the invention take a fuzzy safety box method as a basic framework, adopt the number of ridge lines between the position of a fingerprint detailed point and the adjacent three detailed points thereof as verification characteristics, extract stable reference points from fingerprints and blocking direction fields of the peripheral specific neighborhoods thereof as assisted characteristics, use a method based on the direction field mutual information to evaluate the coincidence degree of the fingerprint registration, and search the optimum parameters by a searching strategy from roughness to refinement, namely, the optimum parameters are the registration parameter for the template fingerprint and the querying fingerprint. The registration parameters are used to convert the querying fingerprint characteristics, and encryption and verification operation are carried out on the template fingerprint by using the converted querying fingerprint verification characteristics.

The present application relates to an image registration method, comprising: selecting a registration source image {pi} and a registration target image {qj}; applying a random perturbation to {pi} in accordance with a preset random perturbation control parameter σ so that it is deformed to obtain {pi′}, and obtaining a set of closest points, i.e., {gj′}, on {qj} corresponding to points on {pi′}; performing an iterative operation on {pi′} and {qj′} in accordance with a preset initial coordinate transformation H0 to obtain a coordinate transformation {H1}, 0<I<=L, between {pi′} and {qi′}, and calculating an average distance {E1} corresponding to the coordinate transformation {H1} in accordance with the coordinate transformation {H1}; judging magnitudes of the average distance {E1} and a preset ideal average distance Ex, and terminating the registration when the average distance {E1} is smaller than or equal to the preset ideal average distance.

The invention discloses a medical image registration method based on quantum evolutionary computation and B spline conversion, and mainly aims to solve the problems that in the prior art, the matching degree is low, and the registration preprocessing process is complex. The method includes the steps that images are read in, and relevant parameters are initialized; quantum evolutionary computation populations are generated at random; registration parameters corresponding to individuals in the populations are calculated; three times of B spline function conversion is conducted on the floating images according to the registration parameters to acquire the converted floating images; normalization mutual information values between the converted floating images and a reference image are computed, a local generation optimal registration parameter is looked for, and quantum revolution door update is conducted on all the individuals in a local generation population; quantum regrouping operation is conducted through the updated population; circulation conditions are controlled, and the registration images and the registration parameters are output. Compared with a method based on a genetic algorithm, a better registration result can be acquired, registration time is shortened, and the method can be used for registering the medical images.

The invention discloses a film mark-point-free automatic positioning method. The method comprises the following steps: loading film dosage data and plan data corresponding to the film dosage data; obtaining data having the same pixel dimension as the plan data through interpolation; selecting a focus area; searching for an optimum coupling area of the focus area in the plan data through translation and rotation operation and obtaining a translation parameter and a rotation angle; and performing, according to the translation parameter, the rotation angle and the zoom information of the interpolation, directional operation on the film dosage data and obtaining data coupling to the plan data. By using the method provided by the invention, errors caused by manual operation in a registering process can be effectively prevented.