32 results about "Morphological model" patented technology

The invention provides a large-amplitude face straightening method by means of an adversarial network and a three-dimensional morphological model. The main content of the large-amplitude face straightening method by means of an adversarial network and a three-dimensional morphological model includes a reconstruction module, a generation network and classification module, and an identification module. The large-amplitude face straightening method by means of an adversarial network and a three-dimensional morphological model includes the steps: a generator generates a forward front face image by taking a non-forward front face image as input, and at the same time a classifier tries to determine whether the image is a real image and utilizes the fed back information to promote the image generated from the generator to be more close to the real image, and at the same time an identification engine is used to maintain the original identity characteristics in the input image. The large-amplitude face straightening method by means of an adversarial network and a three-dimensional morphological model can process the non forward face, especially a large-amplitude deflected face image, can provide a generation network and a morphological model to straighten the face, and can greatly improve the effect of face identification and straightening at the same time.

An automated method for the generation of (i) human models comprehensive of shape and joint centers information and / or (ii) subject specific models from multiple video streams is provided. To achieve these objectives, a kinematic model is learnt space from a training data set. The training data set includes kinematic models associated with corresponding morphological models. A shape model is identified as well as one or more poses of the subject. The learnt kinematic model space and the identified shape model are combined to generate a full body model of the subject starting from as few as one-static pose. Further, to generate a full body model of an arbitrary human subject, the learnt kinematic model space and the identified shape model are combined using a parameter set. The invention is applicable for fully automatic markerless motion capture and generation of complete human models.

The invention relates to an image segmentation method and system. The method comprises the following steps: initially positioning medical images to obtain a positioning area; preprocessing the positioning area to obtain a target area, wherein the target area contains a nodular area and a background area, and the nodular area consists of a solid area and a surrounding area which surrounds the solid area; processing the target area in combination with a Gaussian mixture model to obtain a probability graph of the target area; processing the probability graph according to a morphological model to determine the solid area and the surrounding area of a nodule; and fusing the solid area and the surrounding area to obtain an image segmentation result. According to the method and system provided by the invention, different types of nodules can be accurately segmented, and the follow-up diagnostic analysis can be effectively improved.

The invention discloses a method for identifying signals. The method comprises the following steps: setting types of the signals to be identified, and constructing a template corresponding to each type of signals according to a morphological model of a power spectrum of each type of signals; extracting the power spectrum of the received signals, and calculating the morphological envelope of the power spectrum of the received signals, wherein the morphological envelope of the power spectrum is called envelope for short; respectively calculating the similarity between the envelope of the received signals and each template; and comparing each similarity, and identifying the received signals. By using the method, the problem of low signal identification rate under the condition of low signal-to-noise ratio in the prior art is solved.

The invention relates to the tree morphological visual model simulation model and image correction of close range photography measurement, in particular to a method for rapidly acquiring tree morphological model parameters based on photogrammetry, and belongs to the technical field of a computer. The method comprises the following steps of selecting a tree morphological model in form of a power function, shooting and carrying out auxiliary measurement of tree images, correcting the tree images, and matching the model with the images. Compared with the prior art, the method is specially formulated aiming at the requirements of exactly acquiring peripheral contour information of two-dimensional trees; the method has low requirements to the equipment; and the auxiliary measurement is simple and easy to implement. Thus, the method is good for acquiring field data and popularizing; the selected tree morphological model comprises key indexes, namely crown indexes in the tree morphology; and the method is very good for describing the tree morphology and unifying different tree models.

The invention discloses a database system for quantitative analysis and shape simulation of deep-water channel configuration based on seismic data. The database system comprises a data input module, adata collation module, a data storage module, a screening module, a data output module, a graphics generation module and a comparison module, wherein the data input module is used for inputting three-dimensional seismic coordinates and geological background information data of a work area; the data collation module is used for dividing data into various data types related to deep-water channel configuration; the data storage module is used for storing various data types in a database with a hierarchical structure separately; the screening module is used for screening deep-water channels developed in similar sedimentary environments; the data output module is used for outputting screened characterization parameters or geometric description information; the graphics generation module is used for generating charts or morphological models based on various data stored in the database; and the comparison module is used for calculating the morphological similarity of the model by means of fuzzy model identification. The database relevant to deep-water channel configuration units is established by programming in order to collect and collate quantitative and qualitative data of different configuration units, and carry out quantitative analysis and morphological simulation on the water channel configuration.

The invention relates to an optimum selection method of a wild parent of an autumn shrimp of Marsupenaeus japonicus and a fundamental group morphological model obtaining method. Marsupenaeus japonicus summer wild individual shrimps having the body types meeting a morphological model equation of a Marsupenaeus japonicus autumn shrimp high-quality wild parent selection fundamental group are selected into an autumn shrimp high-quality wild parent fundamental group; after artificial ripening promotion and culture, female shrimps that have characteristics of bright and clean body surfaces, bright and transparent shells, bright body colors, complete appendages, healthy states with no disease and no damage, quick reaction, good gonad development, and long body lengths larger than or equal to 16cm are selected as Marsupenaeus japonicus autumn shrimp breeding parents. Therefore, the overwintering breeding survival rate of the Marsupenaeus japonicus autumn shrimps is effectively improved.

The invention discloses a three-dimensional morphological virtual modeling method of corn silk. The method comprises the following steps: observing the morphological structure of core silk and determining characteristic parameters of three-dimensional morphological of corn silk; establishing a three-dimensional virtual morphological model according to the characteristic parameters, wherein respectively carrying out modeling on silk axis, cluster stalk, spike stalk, bract and kernel. The three-dimensional morphological virtual modeling method combines the modern computer graphics technology andthe modern algricultural science and production process, can realize three-dimensional display of silk on the computer for different corn varieties by setting model parameters, and has stronger realistic effect. The model has strong controllability and easy operation and can meet the requirements of digital design of the corn silk.

The invention discloses a heart soft tissue three-dimensional reconstruction method based on sparse principal component analysis. According to the method, first, a traditional high-complexity thin-plate spline model is used to extract three-dimensional historical morphological data of a region of interest in historical image data; then relative coordinate processing and equalization processing areperformed on the extracted three-dimensional morphological data of the region of interest; principal component analysis is performed on the processed data, and sparse principal components are extracted; and last, the obtained sparse principal components are used to establish a new morphological model, model parameters are obtained from a currently acquired heart surface stereoscopic image throughstereoscopic matching, and then three-dimensional morphology of a current heart surface target region is reconstructed.

The invention provides a method and device for obtaining vascular pressure difference. The method comprises steps that anatomical data of a portion of a vessel segment is received, and a geometric model of a target vessel is obtained according to the anatomical data; a blood flow model of the target vessel and the blood flow velocity V of the target vessel are obtained according to the anatomicaldata in combination with individual data; the geometric model is pre-processed to establish a cross section morphological model, a morphological difference function f(x) of a target vessel lumen is calculated, based on the morphological difference function f(x) of the target vessel lumen and the blood flow velocity V, a pressure difference value deltaP at any two positions of the target vessel iscalculated. The method is advantaged in that through introducing the morphology concept, influence of blood vessel morphology and the like on the vascular pressure difference is clarified, and therebycalculation accuracy of the vascular pressure difference is improved.

An automated method for the generation of (i) human models comprehensive of shape and joint centers information and / or (ii) subject specific models from multiple video streams is provided. To achieve these objectives, a kinematic model is learnt space from a training data set. The training data set includes kinematic models associated with corresponding morphological models. A shape model is identified as well as one or more poses of the subject. The learnt kinematic model space and the identified shape model are combined to generate a full body model of the subject starting from as few as one-static pose. Further, to generate a full body model of an arbitrary human subject, the learnt kinematic model space and the identified shape model are combined using a parameter set. The invention is applicable for fully automatic markerless motion capture and generation of complete human models.

A method for create a morphological model of a rehabilitation brace based on a medical image technology comprises seven steps. There are three advantages in the method of establishing the morphological model of rehabilitation brace based on the medical imaging technology. First, it provides a reference scheme for the model design process of clinical lower limb rehabilitation brace. Secondly, three-dimensional image registration technology can accurately display the size and distribution of the morphological deviation between the heterogeneous, which can provide guidance for the clinical rehabilitation brace manufacturing process. Thirdly, it provides important data support for the future large database construction of intelligent and personalized rehabilitation brace matching.

The invention provides a digital rock physical model construction method and device. The method comprises steps that boundary conditions of a deposition area, the position and the particle size of deposited mineral particles in the deposition area and the initial position and the particle size of to-be-deposited mineral particles are acquired; according to the boundary conditions of the depositionarea, the position of the deposited mineral particles and the initial position of the to-be-deposited mineral particles, the termination position of the to-be-deposited mineral particles in the deposition area after at least one collision between the to-be-deposited mineral particles and the deposited mineral particles is determined; according to the particle size and the termination position ofthe to-be-deposited mineral particles and the position and the particle size of the deposited mineral particles, the position and the particle size of the updated deposited mineral particles are determined; a morphological model of the rock mineral particles in the deposition area is established based on the position and the particle size of the deposited mineral particles after update and the boundary conditions of the deposition area. The method is advantaged in that authenticity and validity of the digital rock physical model can be effectively improved, and instructive advices can be provided for petroleum exploration.

The invention discloses an accurate three-dimensional cell morphology recovery method based on a depth change point spread function, and aims to solve the problem of low accuracy of constructed cell models caused by inconsistency of actual scanning step length and a preset value and reduction of image quality in an imaging process of a laser scanning confocal microscope. A ray tracing method is used to calculate and obtain an actual scanning step length, point spread functions of a detection plane under different detection depths are calculated layer by layer according to the actual scanning step length, and the point spread functions of different detection depths are used to carry out deconvolution operation on cell layers of corresponding depths so as to improve the quality of an original image. The accuracy of cell construction is improved in the image plane direction, the images after deconvolution are spatially spliced according to the actual scanning step length, and the accuracy of cell construction is improved in the scanning stepping direction. According to the method, a more accurate three-dimensional cell morphological model can be constructed, so that the method is applied to the field of biomedical research and has very important theoretical significance and practical application value.

The embodiment of the invention provides a crop plant phenotype analysis method and device, and belongs to the technical field of agricultural informatization. The method comprises the following steps of acquiring the plant morphological data of crop plants; inputting the plant morphological data into a crop plant morphological model, and outputting the stalk morphological data and the missing leaf morphological data of the crop plant. Due to the fact that the stem morphological data and the missing leaf morphological data of the crop plants can be directly obtained through the crop plant morphological model on the scale of the single plant, compared with a three-dimensional reconstruction mode, the workload is smaller, the efficiency is higher, and the high-throughput phenotypic analysis can be conveniently achieved. In addition, only the external leaves of the wheat plant three-dimensional model constructed by the three-dimensional reconstruction methods, such as multi-view imaging. three-dimensional laser radar, etc., has the relatively higher resolution ratio, the point cloud noise of stalks and leaves close to the center of a plant is high, and the accurate three-dimensional reconstruction is difficult to achieve, but the accurate three-dimensional reconstruction and phenotypic analysis can be achieved through the embodiment of the invention.

The invention provides a method and a device for establishing a wheat growth form model. The method is used for establishing a wheat growth form model in combination with a wheat growth model. The method comprises the steps of according to meteorological parameters in a wheat growth model, dry matter weight of the wheat organs and morphological characteristic relationship between the two and the wheat organs establishing a morphological characteristic parameter model, obtaining field test data including meteorological data, wheat organ characteristic data and wheat organ dry matter weight data, fitting to obtain numerical values of model parameters in the morphological characteristic parameter model according to the field test data and the morphological characteristic parameter model, anddetermining a wheat growth morphological model. According to the wheat growth morphological model and the parameterized geometric modeling method, the wheat growth model is constructed through biomass, the constructed model is higher in mechanology and universality, and the constructed geometric parameter model of wheat leaves, leaf sheaths and stems is better in reliability and practicability. The wheat growth morphological model and the parameterized geometric modeling method are used for constructing the wheat leaf, leaf sheath and stem geometric parameter model.

The invention provides a calculation method and device for obtaining blood vessel pressure difference based on individual gender. The calculation method for obtaining blood vessel pressure difference based on individual gender includes: receiving anatomical data of a part of blood vessel segments and establishing a geometric model and a cross-sectional shape model, and calculating the shape difference function f(x) of the lumen of the target blood vessel; according to the anatomical data Combined with individual specific data, the blood flow model and the first blood flow velocity V 0 ; and through the correction parameter ω to the first blood flow velocity V 0 Perform correction to obtain the second blood flow velocity V; based on the morphological difference function f(x) of the lumen of the target blood vessel and the second blood flow velocity V, calculate and obtain the pressure difference value ΔP at any two positions of the target blood vessel. The calculation method and device for obtaining vascular pressure difference based on individual gender provided by the present invention improve the accuracy of vascular pressure difference calculation by introducing the concept of morphology and the correction parameter ω related to gender information.

The invention relates to the tree morphological visual model simulation model and image correction of close range photography measurement, in particular to a method for rapidly acquiring tree morphological model parameters based on photogrammetry, and belongs to the technical field of a computer. The method comprises the following steps of selecting a tree morphological model in form of a power function, shooting and carrying out auxiliary measurement of tree images, correcting the tree images, and matching the model with the images. Compared with the prior art, the method is specially formulated aiming at the requirements of exactly acquiring peripheral contour information of two-dimensional trees; the method has low requirements to the equipment; and the auxiliary measurement is simple and easy to implement. Thus, the method is good for acquiring field data and popularizing; the selected tree morphological model comprises key indexes, namely crown indexes in the tree morphology; and the method is very good for describing the tree morphology and unifying different tree models.

The invention discloses a three-dimensional reconstruction method of cardiac soft tissue based on sparse principal component analysis. Firstly, the traditional high-complexity thin-plate spline model is used to extract the three-dimensional historical shape data of the region of interest in the historical image data; The three-dimensional morphological data of the region are processed with relative coordinates and averaged; the processed data is subjected to sparse principal component analysis to extract sparse principal components; finally, a new morphological model is established with the obtained sparse principal In the stereo image, the model parameters are obtained through stereo matching, and then the three-dimensional shape of the current target area on the heart surface is reconstructed.

The invention relates to an optimum selection method of a wild parent of an autumn shrimp of Marsupenaeus japonicus and a fundamental group morphological model obtaining method. Marsupenaeus japonicus summer wild individual shrimps having the body types meeting a morphological model equation of a Marsupenaeus japonicus autumn shrimp high-quality wild parent selection fundamental group are selected into an autumn shrimp high-quality wild parent fundamental group; after artificial ripening promotion and culture, female shrimps that have characteristics of bright and clean body surfaces, bright and transparent shells, bright body colors, complete appendages, healthy states with no disease and no damage, quick reaction, good gonad development, and long body lengths larger than or equal to 16cm are selected as Marsupenaeus japonicus autumn shrimp breeding parents. Therefore, the overwintering breeding survival rate of the Marsupenaeus japonicus autumn shrimps is effectively improved.

The invention relates to a modeling method, in particular to a modeling method for a skull morphological model of a head of a 3-10-year-old child, which comprises the following steps: acquiring a headskull morphological and thickness coordinate point set by using Mimics, Geomagic and MATLAB software; simplifying the coordinate point set by adopting a principal component analysis method, and establishing a regression relationship between the principal component score matrix and the age and head perimeter; A child head skull shape and thickness model is established through medical CT images, statistical analysis and the like, and necessary data support is provided for child head research.

The invention provides a method for acquiring pressure difference based on smoking history information. The method includes the steps: receiving anatomic data and a geometric model of part of blood vessel sections; acquiring the blood flow velocity V of a blood flow model of an area-of-interest according to the anatomic data and individual specificity data; building a morphological model of the cross section of each position between a near-end terminal point and a far-end terminal point of a target blood vessel; calculating a morphological difference function f(x) of a lumen of the target blood vessel; calculating a first pressure difference delta P0 based on the morphological difference function f(x) and the blood flow velocity V; modifying the first pressure difference delta P0 based onthe smoking history information of the patient to obtain a second pressure difference delta P. Based on the morphological difference function f(x), the influence of plaque information on the blood vessel pressure difference is defined, blood viscosity u and deviation rectifying parameters k are introduced based on the smoking history information of the patient to modify the first pressure difference delta P0 of the patient, the second pressure difference delta P is acquired, and pressure difference calculation accuracy is improved.

The invention discloses an on-line fitting system, which comprises a human body scanning and collecting unit for collecting images of a user from different directions and scanning a human body to obtain figure data; a model preliminary generation unit which is connected with the human body scanning acquisition unit, and synthesizes a preliminary morphological model of the user according to a preset algorithm by utilizing the image and figure data acquired by the human body scanning acquisition unit; an input modification unit which is connected with the model preliminary generation unit, so that a user inputs representation data, and modifies the preliminary form model generated by the model preliminary generation unit according to the representation data input by the user, thereby obtaining a form refinement model with a higher matching degree with a real person form; a big database unit which is used for storing a clothes model; and a calling synthesis unit which is connected with the big database unit and is used for calling the clothes model in the big database unit by a user. The online fitting system can truly show a fitting effect to a fitting person, and can bring a real fitting feeling to a user to the maximum extent.

The invention discloses a bone destruction degree feature extraction system, which includes a joint statistical morphological model building module, a pathological morphological model matching module, and an abnormality detection and analysis module, wherein the joint statistical morphological model building module is used to construct a joint statistical morphological model, Extract specific morphological statistical features for each piece of bone; the pathological morphological model matching module is used to find a model instance closest to the pathological sample in the joint statistical morphological model, and limit the model instance to three times the standard of the average morphological model Within the difference, the influence of normal anatomical variation is eliminated, and the pathological changes are highlighted; the target information used by the abnormality detection and analysis module is the degree of osteogenesis and osteoclastion of each bone, by calculating the normal The negative distance realizes quantitative evaluation of osteoclast and osteoblast. In this way, the quantitative calculation of bone destruction evaluation indicators can be realized, and reference information can be provided for the evaluation of the position and degree of bone destruction in patients.

The invention provides a method for simulating corneal postoperative conditions, the steps of which include the following steps: step 1, establishing a basic eyeball model through finite element modeling software; step 2, inputting specific object corneal parameters on the basis of the basic eyeball model to establish individual eyeballs Morphological model; step 3, input the individual corneal material properties, and prepare a personalized whole-eye eyeball model; step 4, prepare the ablation surgery marking model by inputting the cutting depth parameters and cutting range parameters in the surgical plan simulation software; step 5, in the ablation surgery On the basis of the marked model, construct an ablation surgery model without intraocular pressure; step 6, remove the corneal tissue according to the position and depth of the marked ablation contour on the ablation surgery model without intraocular pressure, and obtain the postoperative model; step 7, according to the optical path tracing The reference data of postoperative corneal refractive power is obtained through trace method calculation and simulation to evaluate the operation effect and realize the accurate prediction of corneal postoperative situation of the three operation schemes.

The invention discloses a large-scale crop phenology extraction method based on a morphological model method. Obtain the multi-year vegetation index time series curve in the local geographical area; use the double logistic function fitting method to smoothly fit the vegetation index time series curve; use the reference curve and reference point of the agricultural gas station to establish a morphological model; use transformation to fit the morphological model; use The optimal scaling parameter is calculated to obtain the extraction value of the phenological period in the local geographical area. The method of the invention utilizes the macroscopic characteristics of the curve, can reduce the influence of local fluctuation and noise of the curve, and has better extraction accuracy; and can simultaneously extract each phenological period of the crop, without setting different metric standards, and can Applied to the extraction of large-scale crop phenology.

The invention relates to an image segmentation method, a pulmonary nodule detection method and a computer-aided detection system. The image segmentation method comprises the following steps: a lung CT image is acquired; initial positioning is carried out on the CT image to obtain one or more positioning areas; a filter is adopted to enhance the positioning areas to acquire a target area; the target area processed based on a clustering model is judged, if a first judgment is met, a real region for the nodule is acquired; the target area processed based on a morphological model is judged, if a second judgment is met, a surrounding region for the nodule is acquired; and the real region and the surrounding region are fused to acquire an image segmentation result. According to the method and the system provided by the invention, different types of nodules can be accurately segmented, and follow-up diagnostic analysis can be effectively improved.

The invention discloses a database system for quantitative analysis and morphological simulation of deep water channels based on seismic data, including: a data input module for inputting the three-dimensional coordinates of the earthquake and geological background information data of the work area; a data sorting module for dividing the data into deep water channels Various data types related to channel configuration; data storage module, which stores various data types in a database with hierarchical structure; screening module, which screens out deep water channels developed in similar sedimentary environments; data output module, which filters Output the characterization parameters or geometric description information; the graph generation module generates graphs or morphological models based on various data stored in the database; the comparison module uses the fuzzy model to identify the morphological similarity of the calculation model. The invention utilizes programming to establish a database related to configuration units of deep water channels to record and organize quantitative and qualitative data of different configuration units, and perform quantitative analysis and morphological simulation on channel configurations.

This application relates to a method for reconstructing the shape of a deep-sea near-bottom towed multi-channel seismic receiving array, which belongs to the technical field of marine geophysical exploration. The actual observation values ​​of the direct wave and seabed reflection wave travel time, the seabed terrain slope angle and seawater sound velocity corresponding to the location of the excitation point of each shot set; Characterize the displacement and vertical offset; obtain the theoretical calculation expressions of the travel time of the direct wave and the seabed reflection wave received by each shot collection point; construct the geometry model parameter inversion of the deep-sea near-bottom towed multi-channel seismic receiving array. The objective function of the derivation; reconstruct the relative position relationship between the deep-tow seismic source and the receiver point. The application can improve the inversion accuracy of deep-tow seismic source and receiver point positions, and improve the quality of deep-tow multi-channel seismic imaging sections.

The invention discloses a method for identifying signals. The method comprises the following steps: setting types of the signals to be identified, and constructing a template corresponding to each type of signals according to a morphological model of a power spectrum of each type of signals; extracting the power spectrum of the received signals, and calculating the morphological envelope of the power spectrum of the received signals, wherein the morphological envelope of the power spectrum is called envelope for short; respectively calculating the similarity between the envelope of the received signals and each template; and comparing each similarity, and identifying the received signals. By using the method, the problem of low signal identification rate under the condition of low signal-to-noise ratio in the prior art is solved.