354 results about "Head model" patented technology

The invention discloses a transcranial electrical stimulation intracranial vector electric field detection device and method, and belongs to the transcranial electrical stimulation intracranial electric field detection technology, the device inputs an individual head model in a stimulation system, sets a target spot and calculates transcranial electrical stimulation parameters; the implantation position and angle of the vector detection electrode are planned through the navigation system, electrical stimulation is applied to a brain target area through the stimulation electrode, and intracranial electric signals are collected in real time through the vector detection electrode; performing high-precision amplification on the acquired weak signal, and converting the weak signal into a digital signal; the collected electric field signals are calculated and processed in real time, and electric field distribution generated by transcranial electrical stimulation in the brain is analyzed and visualized; according to the method, time domain and frequency domain analysis is carried out on electric signals before and after stimulation, the influence of stimulation on neural signals in a detection area is evaluated, and high-precision measurement of the intensity and direction of a three-dimensional electric field is achieved through the design of a high-precision and multi-layer vector detection electrode.

The invention relates to the technical field of medical instruments, in particular to a high-precision time interference non-invasive brain deep electrical stimulation method, equipment and system, and solves the problems that two stimulation electrodes are used for each pair of current in traditional time interference-based non-invasive brain deep electrical stimulation, so that the current distribution is relatively dispersed, and the stimulation focusing performance is relatively poor. According to the method, high-precision 4 * 1 electrode configuration is used, two electrodes for applying a stimulation current are changed into a 4 * 1 electrode form, the flowing path of the current is effectively limited, the current is more focused, and the focusing performance of the interference current is further improved. According to the device, high-precision current control is realized through the anti-phase output circuit, the 4 * 1 voltage division circuit, the constant current source module and the impedance detection module; the system supports forward and reverse electrode position optimization, a personalized real finite element head model simulation system based on MRI data is used for guiding practical application, the focusing performance of each pair of current in a target brain region is improved, and therefore the overall stimulation precision in the target brain region is improved.

In various examples, head pose prediction for automotive occupant sensing systems and applications is presented. The systems and methods described herein provide for a machine learning model trained using a dataset that comprises ground truth head pose data computed using a registered head model of a training subject. While operating a vehicle, one or more cameras and a depth sensor capture synchronized images of the training subject. To compute a ground truth 3D head pose, angular deviations between a 3D point cloud and the registered head model may be computed to obtain a 3D ground truth head pose measurement. Using an extrinsic calibration transform, the head pose measurement may be mapped into the sensor coordinate frame. Training samples may be produced for training the machine learning model that comprise an optical image frame and the head pose measurement transposed into the frame of reference for that optical image frame.

The invention relates to the technical field of neuroscience and medical treatment, in particular to an individual head model dynamic optimization method and system based on transcranial alternating current stimulation. According to the method, the three-dimensional initial individual head model is partitioned, the parameters are set, the real-time electrical stimulation parameters and the real-time electroencephalogram information of the individual head under the stimulation effect are synchronously obtained, and then the parameters of the initial head model are adjusted according to the real-time data; then, according to the obtained data, simulating and calculating electrophysiological change data of the three-dimensional initial individual head model under the action of initial transcranial alternating current stimulation, and based on a focusing index value and a uniformity index value calculated through the electrophysiological change data, optimizing dynamically-adjusted head model parameters, and finally, calculating a focus index value and a uniformity index value of the three-dimensional initial individual head model under the action of initial transcranial alternating current stimulation. The information entropy difference value obtained through calculation of the optimized head model parameters and the real-time electroencephalogram information is compared with a preset difference threshold value, a clear termination condition is provided for optimization of the head model parameters, and therefore dynamic optimization of the individual head model is achieved.

Disclosed in the present invention is a Gaussian mixture shape method applicable to dynamic modeling of the head of a human body. A Gaussian mixture shape is composed of a neutral Gaussian model representing a neutral expression and a plurality of basic expression Gaussian models corresponding to a plurality of basic expressions. Each Gaussian model comprises a group of three-dimensional Gaussian functions having information such as position, shape, and color. A human face and head model of any expression is efficiently generated by linearly mixing the neutral Gaussian model and the basic expression Gaussian models. The method comprises training image acquisition and preprocessing, Gaussian mixture shape training, and human head animation generation. The present invention has higher drawing quality, can retain more high-frequency details of the head of the human body, and improves the sense of reality and fidelity of drawing; and the present invention has higher operating speed which is more than five times higher than that of the current most advanced method. Therefore, the present invention can be widely applied to movies, animation production, online games, teleconferences, etc., and has high practical value.

The invention discloses a nerve regulation and control intervention system based on time domain interference electrical stimulation, and the system comprises a magnetic resonance image registration module which receives and registers an imported magnetic resonance image, and obtains a to-be-processed MRI image; the head model tissue segmentation module is used for performing head model tissue segmentation on the to-be-processed MRI image; the stimulation simulation module is used for performing stimulation simulation based on the tissue segmentation structure to obtain a front lead field matrix; the intervention coordinate selection module is used for determining a target brain region corresponding to a target spot brain region coordinate position needing stimulation intervention and putting the target brain region into a running sequence; the electrode position arrangement optimization module is used for verifying, optimizing and finely adjusting the electric field intensity of the target brain region under different electrode arrangement schemes to obtain an electrode configuration parameter result and displaying the electrode configuration parameter result; the parameter importing module is used for receiving the selected electrode configuration parameters and imported preset electrical stimulation parameters; and the electrical stimulation module is used for performing electrical stimulation on the target brain region of the target object. According to the application, the accuracy and the effectiveness of nerve regulation and control treatment can be realized.

The invention discloses a brain power supply imaging method and system based on a physical information neural network. The method comprises the following steps: establishing a three-layer boundary element method head model based on a Maxwell equation set under quasi-static approximation and calculating a lead matrix; carrying out filtering, artifact suppression and standardization processing on the multi-channel electroencephalogram signals; inputting the standardized signal into a neural network encoder to extract a time sequence feature tensor, estimating the source current density based on the feature tensor, and performing forward calculation by using the lead matrix to obtain a predicted scalp potential; constructing a mixed loss function consisting of a data fidelity item, a physical consistency item and a physical heuristic regularization item, and performing network training and optimization by taking minimization of the loss function as a target; and when the training meets a convergence condition, outputting a visualization result of the brain endogenous current density spatial distribution. According to the method, electromagnetic physical constraints are introduced, so that the physical consistency, interpretability and stability of electroencephalogram inversion are improved.

The invention relates to a privatized large model fine tuning training method and device based on split learning, and the method comprises the steps: splitting a pre-trained large model into a head model segment, a middle model segment, and a tail model segment; splitting the original data into private data and non-private data; according to the privacy data and the non-privacy data, cooperatively training a head model segment and a tail model segment in the client and a first intermediate model segment and a second intermediate model segment in the server to obtain a target large model; wherein in the cooperative training process, a privacy intermediate activation amount which is transmitted between a client and a server and is related to privacy data is subjected to homomorphic encryption; the target large model outputs a document reply about a particular topic. According to the method and the device, the problem that the risk of speculating or recovering the original data through model inversion attack exists due to the fact that the intermediate activation amount needs to be transmitted to the server in the prior art is solved; the privacy data is protected, and the risk of privacy data leakage is reduced.

The invention discloses an epilepsy region positioning method and system based on brain power supply imaging and a dynamic brain network, and the method comprises the steps: obtaining T1 weighted magnetic resonance imaging data of a user, and constructing an individual three-dimensional head model through a boundary element method; acquiring scalp electroencephalogram data of a user, and preprocessing the scalp electroencephalogram data; based on an individual three-dimensional head model, performing inverse problem solving on the preprocessed scalp electroencephalogram data by using a standardized low-resolution brain power source imaging algorithm to obtain source current density signals of 68 brain regions; decomposing into six frequency bands, calculating the power spectrum density of each brain region and carrying out normalization processing, and screening effective frequency bands; based on the source current density signals of the 68 brain regions of the effective frequency band, information flow directions and intensities of different brain regions are calculated by adopting a directional transfer function method, a directional transfer function matrix of the effective frequency band is formed, and a directed brain network is constructed; and calculating a graph theory index and / or an epilepsy index of each brain region, carrying out maximum value normalization analysis, and determining an epilepsy region positioning result.

The invention discloses a digital human 3D face reconstruction method, computer equipment and a readable storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: obtaining a plurality of different expression models under an ARKit framework, and creating a basic 3D head model based on 3D modeling software; fitting each expression model by using the basic 3D head model to determine multiple groups of expression parameters and shape parameters, and creating a target 3D head model based on the multiple groups of expression parameters and shape parameters and the basic 3D head model; and performing 3D face reconstruction on the face image based on the target 3D head model. According to the method, the 3D face reconstructed based on the face image can be richer in shape and expression expression and is close to a real person.

The invention discloses a method, device and equipment for generating a head model, a storage medium and a program product. The method is characterized by comprising the following steps of: receiving and verifying original three-dimensional T1 weighted image data and an original three-dimensional grid model; performing tissue segmentation based on the verified three-dimensional T1 weighted image data and the original three-dimensional T1 weighted image data to obtain a cerebral grey matter mask, a cerebral white matter mask, a scalp mask and a skull mask; performing topological structure repair on the brain grey matter mask and the brain white matter mask to obtain a brain mask; respectively converting the scalp mask, the skull mask and the brain mask into a scalp mesh model, a skull mesh model and a brain mesh model; calculating and applying a spatial transformation matrix from the verified three-dimensional grid model to the scalp grid model, and performing spatial registration and fusion on the scalp grid model, the skull grid model and the brain grid model to obtain a head model; the method has the advantages that the efficiency, precision and robustness of head model generation are effectively improved, and the method has good cross-platform deployment capability.

The invention provides a brain low-frequency conductivity tensor modeling method, a brain low-frequency conductivity tensor modeling system, a medium and electronic equipment. The method comprises the following steps: acquiring a phase image of a fast spin echo sequence of a brain; acquiring a high-frequency conductivity image of the brain based on the phase image; collecting a diffusion weighted amplitude image of the brain based on a multi-excitation diffusion weighted sequence; based on the diffusion weighted amplitude image, using a multi-interlayer model to obtain diffusion microstructure parameters of the brain; generating a low-frequency conductivity tensor image of the brain based on the high-frequency conductivity image and the diffusion microstructure parameters; generating a digital head model based on the brain magnetization preparation fast gradient echo sequence; and generating a craniocerebral low-frequency conductivity tensor model based on the low-frequency conductivity tensor image and the digital head model. According to the brain low-frequency conductivity tensor modeling method and system, the medium and the electronic equipment, more accurate target point planning is provided for the tTIS.

The invention discloses a three-dimensional head model generation method based on Gaussian point cloud reconstruction of forward propagation, and the method comprises the steps: compressing an input image to a potential space, obtaining a potential state, and extracting the identity information of the input image at the same time in a stage of generating a multi-view image; de-noising is carried out through a de-noising U-Net with a space and time attention module, and a multi-view image is generated through a VAE decoder and is used for training a weight fine tuning network; in the Gaussian reconstruction stage, the world coordinate origin and the light direction of the camera where each pixel in the multi-view images generated by the video diffusion model is located are calculated, the multi-view images are spliced in the channel dimension through the corresponding light embedding and light direction obtained through calculation, network input features are formed, and the network input features are used as network input features. An asymmetric Gaussian generation UNet based on LGM network improvement is utilized, a feature map is generated according to a multi-view image generated by a video diffusion model, and channel flattening processing is performed on the feature map to obtain a Gaussian point cloud. According to the invention, a more vivid and lifelike 3D head model with high quality can be generated.

The invention discloses a single-image three-dimensional head reconstruction method and system based on a guide diffusion model. The method comprises the following steps: obtaining a head portrait from a single image to be reconstructed; reconstructing a three-dimensional head model according to the head portrait; performing weak projection mapping on the three-dimensional head model under the predicted camera matrix to the head portrait, and obtaining a three-dimensional head model with front textures by utilizing pixel information of the head portrait; performing cylindrical surface UV expansion on the three-dimensional head model with the front texture to obtain a two-dimensional UV texture map and recording a UV mapping relation; inputting the two-dimensional UV texture map into the guided diffusion network model to predict the texture of the invisible area, and repairing to obtain a complete two-dimensional UV texture map; and then the complete two-dimensional UV texture image is reversely pasted to the three-dimensional head model. According to the method, the three-dimensional head model which is high in fidelity, free of artifacts and rich and vivid in texture details can be reconstructed from individual single portrait images of different ages, genders, races and facial expressions.

According to the neural regulation target positioning method and system based on magnetic resonance and electroencephalogram fusion and the focus brain tissue head model construction method based on deep learning, the individual brain structure can be accurately simulated, and a reliable basis is provided for subsequent analysis; according to the electroencephalogram source imaging method based on deep learning, a brain source can be effectively positioned from complex electroencephalogram signals; a target network accurate positioning method based on source imaging further locks and regulates a target spot and a target brain network. The combination of the three can give full play to the advantages of multi-modal data, achieves more accurate and efficient target spot and target network positioning by means of the powerful ability of deep learning, and provides powerful support for nerve disease diagnosis and treatment.

The invention relates to a heterogeneous human head model construction method and system oriented to multi-physics field simulation, based on real human head T1 weighted magnetic resonance image data, key tissues such as white matter, gray matter and cerebrospinal fluid are accurately segmented, and a three-dimensional geometric model used for generating carrying tissue identification information is researched and developed; converting the model data into a file format conforming to the NASTRAN standard by utilizing an independently developed format conversion program, and completely retaining organization classification information; finally, different tissue areas are automatically recognized through physical field simulation software, corresponding frequency dependent material attributes are distributed, and a finite element model capable of being directly used for simulation is formed. According to the method, seamless integration of the high-precision heterogeneous human head model and the simulation platform is achieved, tedious manual repairing and assignment operation is not needed, the method has the advantages of being high in precision, high in efficiency and good in compatibility, and the authenticity and reliability of biomedical simulation such as transcranial magnetic stimulation and radio frequency coil design can be remarkably improved.

The invention relates to the technical field of photo three-dimensional processing, and discloses a three-dimensional head portrait model generation method based on a single photo, which comprises the following specific steps: photo input processing: firstly, extracting a face region from the single photo by using a face feature point detection algorithm; background segmentation is carried out by using an SAM algorithm, and a face region in the picture is extracted; finally, the depth information of the human face is speculated through a Depth Anything model, and the depth information is used for guiding generation of Gaussian point cloud; according to the three-dimensional head model generation method based on the single photo, the high-precision three-dimensional head model can be generated from the single photo. According to the method, the 360-degree rotating video is firstly generated, and then three-dimensional reconstruction is carried out by utilizing Gaussian splashing, so that the finally generated 3D head model is highly matched with the input picture, and the similarity of the model is relatively high. Besides, the method can retain abundant detail features, and the Gaussian splashing point cloud can accurately restore personalized features of the human face, such as wrinkles, skin colors, hairlines and the like, so that the model is more real and natural.

The invention discloses a transcranial magnetic navigation body surface visualization and autonomous adjustment system. According to the system, a patient head model is reconstructed through the sMRI technology to obtain a reconstructed head model coordinate system, a real head coordinate system under the view angle of a camera is established through the face recognition technology, and two visual visualization schemes are provided: virtual-real mapping of a target spot is realized through a laser holder; a physician is guided to accurately position the transcranial magnetic coil by using an augmented reality technology, and meanwhile, a visual interface is supported to guide a patient to autonomously adjust the position so as to ensure target attachment. The effectiveness of technologies such as transcranial magnetism can be remarkably improved through accurate target fitting, so that deterioration of nerve diseases is delayed, and the health level and life quality of patients are maintained.

The invention discloses an antenna performance analysis method and device for a mobile terminal, and relates to the technical field of communication, and the method comprises the steps: obtaining different antenna digital models and the working frequency corresponding to each antenna digital model; placing each antenna digital model at different positions of the mobile terminal shell model to obtain a plurality of mobile terminal models of each antenna digital model; according to various mobile terminal position association information, the mobile terminal model is combined with the head model and the hand model, and different simulation scenes are constructed; determining information such as current density distribution of the mobile terminal model in a free space, OTA performance and electromagnetic radiation in different simulation scenes by using an electromagnetic simulation algorithm; training the neural network model by using the simulation result to obtain an antenna performance analysis model; and performing antenna performance analysis by using the antenna performance analysis model. The antenna performance analysis cost of the mobile terminal can be reduced, and the analysis efficiency and the accuracy of the analysis result can be improved.

The invention relates to the technical field of medical instruments, and discloses a positioning method and system for rapid registration in transcranial magnetic stimulation navigation, and the method comprises the steps: S1, data collection: collecting RGBD image data of the face of a patient through a high-precision 3D camera, and obtaining skull 3D data through magnetic resonance imaging and computed tomography; s2, data processing: processing the RGBD image, and obtaining feature point information of the face of the patient and a head 3D model; s3, establishing a transcranial magnetic stimulation point model, mapping the skull 3D data to the head 3D model, and establishing a patient personalized head model and a magnetic stimulation point model.According to the method, efficient and accurate positioning in the transcranial magnetic stimulation treatment process is achieved by integrating the steps of high-precision data collection, space positioning and modeling, path planning, rapid registration, real-time tracking and the like, and the accuracy of the treatment process is improved. And the treatment effect and the patient experience are improved.

The present application relates to the field of imaging effect testing, and discloses a three-dimensional imaging effect testing system and method, a device and a storage medium. The system (100) comprises: a head model (101), a binocular camera (102) provided at a binocular position of the head model (101), and a computing component (103) connected to the binocular camera (102). The binocular camera (102) is used for capturing a three-dimensional visual image (13) presented by a three-dimensional display device (104) to obtain a first left-eye image (111) and a first right-eye image (121). The computing component (103) is used for outputting imaging effect information of the three-dimensional display device (104) on the basis of the first left-eye image (111) and the first right-eye image (121). The solution can evaluate the imaging effect of the three-dimensional display device by means of objective data.

The invention discloses a spatial audio test method, which comprises the following steps: constructing a stereo microphone array, placing the stereo microphone array in a simulation head model, and placing a test earphone outside a simulation head ear canal of the simulation head model; the test earphone plays a spatial audio signal, the stereo microphone array collects the spatial audio signal, and the collected multi-channel spatial audio signal is sent to the main control system; the main control system carries out format conversion on the multi-channel spatial audio signal and transmits the multi-channel spatial audio signal to the host processing system through a USB, the host processing system processes the spatial audio signal after format conversion to obtain a compensated spatial audio signal, and then the compensated spatial audio signal is processed by using a positioning algorithm to obtain an audio signal; and obtaining a test result of the sound source corresponding to the spatial audio signal. The millimeter-level precision test of the earphone spatial audio is realized, the core challenge in the earphone spatial audio test is solved, and a golden standard test platform is provided for the development of a spatial audio algorithm.

The invention relates to the field of computer vision and graphics, in particular to a three-dimensional head model reconstruction method and device and electronic equipment. The method comprises the following steps: acquiring a plurality of head images of the same head, wherein the plurality of head images at least comprise a front face image and a side face image; respectively converting the plurality of head images into images of a Tensor object; acquiring face region images according to the image of the Tensor object, wherein each face region image at least comprises eye image information, nose image information and mouth image information; acquiring face key point data according to the face region image; and establishing a three-dimensional head model according to the face key point data. According to the embodiment of the invention, the combination of a traditional image processing method and a depth algorithm is realized, and the image data is processed by using a depth detection algorithm through a large amount of image data, so that the generated three-dimensional head model is close to a real object, the problems in the prior art are solved, and the accuracy and efficiency of reconstruction of the three-dimensional head model are improved.

The invention discloses a rapidly pre-assembled model product, which is formed by mutually sleeving and assembling a trunk model, a head model, an arm model and a leg model; The head model is arranged on the top of the trunk model; The arm model includes a left arm model and a right arm model, wherein the left arm model is arranged on the left side of the top of the trunk model, and the right arm model is arranged on the right side of the top of the trunk model; The leg model includes a left leg model and a right leg model, wherein the left leg model is arranged on the left side of the bottom of the trunk model, and the right leg model is arranged on the right side of the bottom of the trunk model. The figure model of the invention is formed by mutually sleeving and assembling a trunk model, a head model, an arm model and a leg model, thus canceling the splicing form of the existing model adopting a small block structure, and through the pre-assembled modules, the modules are assembled and spliced through corresponding connecting structures, so as to make the whole model product assembled and spliced more conveniently and quickly.

The invention discloses a novel multi-angle head drop hammer impact device and a test method. According to the device, a steering engine is adopted to drive a clamp to assemble a drop hammer, electromagnetic interference is avoided, friction of a motor shaft is also avoided, and repeatability and accuracy of an experimental result are improved; the supporting arc plate for clamping the head model and the bottom plate are fastened through meshing of teeth in two directions, so that the fastening effect is extremely high, movement is not prone to occurring when the head model bears impact of a drop hammer, the fastening mode is simple and easy to operate, and rapid and repeated testing is facilitated; a movable mounting structure is arranged between the hammer head and the impact plate, so that the position of the hammer head can be finely adjusted in all directions, the position deviation caused by tooth meshing can be conveniently adjusted, in addition, a gel ice and snow plate is arranged around the head model, the actual temperature on the head model is controlled, and the impact effect at the low temperature can be simulated.

The invention relates to the technical field of robot collision test, and provides a human body model for collaborative robot interaction collision test and a control method thereof, and the human body model comprises a head module which comprises a head supporting part and a head model covering the outer layer of the head supporting part, and a three-dimensional force sensor is fixed on the side surface of the head module; the neck module comprises a neck skeleton rotationally connected with the head supporting part, and a six-dimensional force sensor is arranged at the joint; the trunk module comprises a trunk skeleton hinged to the neck skeleton and a thoracic cavity model covering the outer layer of the trunk skeleton, a plurality of three-dimensional force sensors are arranged on the front chest part of the trunk skeleton, and a three-dimensional force sensor and a flexible sensor are arranged on the back part; the big arm skeleton is connected with the trunk skeleton and is provided with a three-dimensional force sensor; the pose adjusting module comprises a horizontal rotating structure and a linear module, and simulates human body steering and gait movement so as to drive the head module, the neck module, the trunk module and the upper limb module to move and collide with the collaborative robot. Omnidirectional capture and analysis of the space collision force are realized.

The invention discloses a 360-degree drivable three-dimensional head portrait method supporting multi-mode generation. The method comprises the following steps: generating a 360-degree head image in combination with a three-dimensional generative adversarial network; three-dimensional Gaussian flattening is attached to a surface patch of the initialized FLAME model, and parameters of the FLAME model are optimized; the vertex displacement of the FLAME model is further optimized, and a refined grid model structure is obtained; using a differentiatable renderer to optimize the texture map; and the shape and size of the three-dimensional Gaussian are freely optimized, so that the three-dimensional Gaussian captures a high-frequency detail region, and finally, a three-dimensional head image which has high rendering quality and driving performance and supports various generation input forms is generated. By combining the three-dimensional generative adversarial network, the FLAME model and the differentiable renderer, a brand new generation framework is provided, and a high-quality and full-view renderable three-dimensional head generation function is realized from multiple input forms.

The invention relates to a virtual character model generation method and device, a storage medium and electronic equipment, and relates to the technical field of computers.The method comprises the steps that target image generation parameters needed for generating a two-dimensional original picture of a virtual character are determined, and the two-dimensional original picture of the virtual character is generated based on the target image generation parameters; constructing an original three-dimensional model corresponding to the virtual character according to the two-dimensional original picture, and extracting a model wiring grid and a model map in the original three-dimensional model; adding a skin pore texture map for the original three-dimensional model according to the model wiring grid and the model map; and generating a target model map according to the skin pore texture map and a basic model map corresponding to the two-dimensional original picture, and rendering the original three-dimensional model based on the target model map to obtain a target three-dimensional model of the virtual character. According to the invention, the generation efficiency of the three-dimensional head model is improved.

The invention relates to the technical field of vehicle collision measurement, in particular to a dummy head displacement measurement method and system based on a sensor and a head model, and the method comprises the steps: marking a camera, a body-in-white backboard reference point and a dummy head key position in a test scene before a test is started, and forming corresponding marking points; recording an initial three-dimensional coordinate corresponding to each mark point through a three-coordinate measuring instrument; parameters of acceleration sensors arranged on the mass center of the body in white and the mass center of the head of the dummy are set; after the simulation collision test is finished, collecting video data shot by the camera and acceleration data recorded by the acceleration sensor; determining displacements of the mark points on the key positions of the head of the dummy in X, Y and Z, and outputting a displacement curve corresponding to the mark points on the key positions of the head of the dummy; and determining the relative displacement of the corresponding dummy mass center in the X, Y and Z directions at the Tmax moment.

The invention provides a head model reconstruction method based on Gaussian splash and a generative model. The method comprises the following steps: fitting a shot face of a person; the expression parameters and the posture parameters of the FLAME head model are optimized; training the 3D Gaussian ball based on a Gaussian splashing algorithm; and obtaining a complete head model reconstruction result based on the generative model. According to the method, the head appearance is subjected to multi-view synthesis through the generative model, a multi-view image which has consistent texture information and surrounds the head by 360 degrees can be inferred, and information loss of the input image in a non-front area is made up; and meanwhile, the multi-view image can also adopt a Gaussian splash algorithm to train the side and back areas of the head model, so that the head model with higher texture precision and more complete structure is reconstructed, and experimental results show that the detail fidelity of the reconstructed model is improved while the completeness of the head structure is ensured.