CN114092634A - Channel and acupoint positioning method and device, electronic equipment and storage medium - Google Patents
Channel and acupoint positioning method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN114092634A CN114092634A CN202111187653.1A CN202111187653A CN114092634A CN 114092634 A CN114092634 A CN 114092634A CN 202111187653 A CN202111187653 A CN 202111187653A CN 114092634 A CN114092634 A CN 114092634A
- Authority
- CN
- China
- Prior art keywords
- human body
- positioning
- dimensional
- model
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 136
- 210000000988 bone and bone Anatomy 0.000 claims abstract description 78
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000001467 acupuncture Methods 0.000 claims description 19
- 238000012545 processing Methods 0.000 claims description 14
- 230000011218 segmentation Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 9
- 210000003813 thumb Anatomy 0.000 claims description 8
- 238000002372 labelling Methods 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 abstract description 33
- 238000005452 bending Methods 0.000 abstract 1
- 230000006870 function Effects 0.000 description 22
- 230000036544 posture Effects 0.000 description 14
- 238000012549 training Methods 0.000 description 12
- 238000004422 calculation algorithm Methods 0.000 description 8
- 239000002775 capsule Substances 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000009826 distribution Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 210000001503 joint Anatomy 0.000 description 4
- 210000004072 lung Anatomy 0.000 description 4
- 210000003414 extremity Anatomy 0.000 description 3
- 230000036541 health Effects 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 208000010392 Bone Fractures Diseases 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 210000002310 elbow joint Anatomy 0.000 description 2
- 210000003127 knee Anatomy 0.000 description 2
- 210000000629 knee joint Anatomy 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000035515 penetration Effects 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 210000002659 acromion Anatomy 0.000 description 1
- 230000037237 body shape Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 210000000245 forearm Anatomy 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 210000004705 lumbosacral region Anatomy 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 238000000554 physical therapy Methods 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000011514 reflex Effects 0.000 description 1
- 210000000323 shoulder joint Anatomy 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000002560 therapeutic procedure Methods 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 210000001364 upper extremity Anatomy 0.000 description 1
- 210000000707 wrist Anatomy 0.000 description 1
- 210000003857 wrist joint Anatomy 0.000 description 1
- 239000008832 zhongfu Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H39/00—Devices for locating or stimulating specific reflex points of the body for physical therapy, e.g. acupuncture
- A61H39/02—Devices for locating such points
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/003—Reconstruction from projections, e.g. tomography
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
- G06T17/205—Re-meshing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Rehabilitation Therapy (AREA)
- Epidemiology (AREA)
- Pain & Pain Management (AREA)
- Physical Education & Sports Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Image Processing (AREA)
Abstract
The invention provides a channel and acupoint positioning method, a device, electronic equipment and a storage medium, wherein the method comprises the following steps: obtaining a single RGB image; preprocessing the single RGB image to obtain a target human body image; acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points; and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings. According to the channel and acupoint positioning method and device, the electronic equipment and the storage medium, the three-dimensional human body model is obtained through modeling based on the single human body image, and the channel and acupoint positioning process is carried out on the three-dimensional human body model by adopting a bone degree bending quantity positioning method, so that the channel and acupoint positioning process is simplified, and the channel and acupoint positioning accuracy is improved.
Description
Technical Field
The present invention relates to the field of medical technology, and in particular, to a method and an apparatus for locating a meridian point, an electronic device, and a storage medium.
Background
People pay more and more attention to body health, meridian therapy is a common health management method, and people hope to achieve the effects of treating diseases and relieving symptoms by massaging meridian points. At present, the meridian and acupuncture points of the human body are mainly identified by referring to a human meridian and acupuncture point model or identifying the meridian and acupuncture points on the human body or identifying the meridian and acupuncture points through self experiences of traditional Chinese medical workers.
The human body is a three-dimensional space body with quite complicated and irregular geometric shapes, the positions of acupuncture points or meridians of each person are different due to the differences of individual shapes, ages, sexes, constitutions and the like, when meridian and acupuncture points are positioned, traditional Chinese medicine workers are required to have years of experience accumulation to find the positions quickly and accurately, inexperienced doctors can be positioned accurately for a long time, and when the positioning is inaccurate, the due physical therapy effect cannot be achieved, unnecessary body injuries can be caused, and the health management effect cannot be achieved.
Disclosure of Invention
The invention provides a channel and acupoint positioning method, a device, electronic equipment and a storage medium, which are used for solving the problems in the prior art.
The invention provides a channel and point positioning method, which comprises the following steps: obtaining a single RGB image; preprocessing the single RGB image to obtain a target human body image; acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points; and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
According to the acupoint positioning method provided by the invention, the preprocessing of the single RGB image to obtain the target human body image specifically comprises the following steps: and carrying out image enhancement processing and/or human body segmentation processing on the single RGB image to obtain the target human body image.
According to the acupoint positioning method provided by the invention, the acquiring of the joint points in the target human body image and the construction of the three-dimensional human body model of the target human body based on the joint points specifically comprise: and acquiring joint points in the target human body image by using an Open Pose method or an Alpha Pose method, and acquiring a three-dimensional human body model of the target human body based on the joint points.
According to the acupoint positioning method provided by the invention, the acupoint positioning is carried out based on the initial reference point and the bone degree cun, and the method specifically comprises the following steps: acquiring the three-dimensional space coordinates of the initial reference points, and taking the initial reference points as reference points; and taking the bone length dimension as a measurement basis of the distance between the channels and the points to obtain the three-dimensional space coordinate of the target channels and points.
According to the acupoint positioning method provided by the invention, the acupoint positioning is carried out based on the initial reference point and the bone degree cun, and the method specifically comprises the following steps: and acquiring the inter-digital width of the thumb in the three-dimensional human body model, correcting the bone degree cun based on the inter-digital width, and performing acupoint location based on the initial reference point and the corrected bone degree cun.
According to the present invention, there is provided a method for acupoint positioning, the method further comprising: when a second positioning method is selected, inputting the three-dimensional human body model into a trained three-dimensional human body positioning model to obtain the three-dimensional human body model for completing acupoint positioning; the three-dimensional human body positioning model is trained by utilizing at least one three-dimensional human body model for completing acupoint labeling.
The present invention also provides a meridian point positioning device, comprising: an acquisition module to: obtaining a single RGB image; a processing module to: preprocessing the single RGB image to obtain a target human body image; a build module to: acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points; a positioning module to: and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of any of the above-mentioned acupoint positioning methods when executing the program.
The invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method for acupoint positioning as described in any of the above.
According to the channel and acupoint positioning method, device, electronic equipment and storage medium, a three-dimensional human body model is obtained through modeling based on a single human body image, and a bone degree folding method is adopted on the three-dimensional human body model to carry out a channel and acupoint positioning process; or training the three-dimensional human body positioning model by using the three-dimensional human body model for finishing the channel and acupoint positioning, and then realizing the channel and acupoint positioning process of the human body by using the three-dimensional human body positioning model; based on the two methods, the channel and point positioning process is simplified, and the channel and point positioning precision is improved.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of the acupoint positioning method provided by the present invention;
FIG. 2 is a flow diagram of a three-dimensional mannequin modeling process provided by the present invention;
FIG. 3 is a schematic view of the bone fracture dimension division of the lumbar region of the back provided by the present invention;
FIG. 4 is a schematic diagram of the lung meridian point location of hand Taiyin provided by the present invention;
FIG. 5 is a schematic structural view of the acupoint positioning device provided in the present invention;
fig. 6 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic flow chart of the acupoint positioning method provided by the present invention, as shown in fig. 1, the method includes:
s110, obtaining a single RGB image;
s120, preprocessing the single RGB image to obtain a target human body image;
s130, acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points;
and S140, when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
It should be noted that, a single RGB image is acquired by using an image acquisition device, and may be acquired by shooting a target human body through a common optical camera or a single lens reflex camera, in order to ensure the integrity of the single RGB image, when the target human body is acquired, it is required to ensure that the complete human body is placed in a view finder frame, and in order to reduce the workload and difficulty of subsequent image processing, a place with less impurities and simple color is selected as a shooting background as much as possible.
The method comprises the steps of preprocessing a single acquired RGB image to obtain a target human Body image, wherein the preprocessing comprises image enhancement (such as image contrast enhancement, image restoration, image definition enhancement, image color enhancement, image illumination adjustment and the like) and human Body segmentation (such as U-Net human Body segmentation, Deeplab V1-Deeplab V3+ human Body segmentation, Body Pix and the like), and the image enhancement is an image processing method which changes an original unclear image into clear or emphasizes certain interesting features, inhibits the uninteresting features, improves the image quality and the information content, and enhances the image interpretation and identification effects. The human body segmentation is to segment the human body from the image background by inputting the acquired image containing the human body into a human body segmentation model, and more precisely, the human body segmentation is to segment different parts of the human body, such as a left foot, a right foot and the like.
Positioning in a target human body image by using a human body joint point positioning method to obtain a corresponding two-dimensional human body joint point, wherein the positioning method can be an Open Pose method, an Alpha Pose method and the like, but is not limited to the two methods listed above; the Open Pose and the Alpha Pose are deep learning-based methods, a human body joint point positioning model is trained by using training data marked with a human body rectangular frame and human body joint point coordinates, a test image containing a human body is input after the model is trained, and then different joint point coordinates on the human body and confidence coefficients of the different joint point coordinates are output. Based on the located two-dimensional human joint points, corresponding SMPL three-dimensional human models, such as SMPLIFY, HMR, etc., can be constructed by a plurality of technical methods.
The SMPL (A Skinned Multi-Person Linear Model) is a parameterized human body deformation Model generated by a statistical learning method through massive real human body scanning data, can represent different postures and body types of a human body through parameter change, and can simulate the protrusion and depression of tissues such as human muscles in the limb movement process, thereby carrying out human body modeling and animation driving of any posture.
It should be noted that the idea of the SMPLIFY method is as follows: the three-dimensional skeleton coordinates of the SMPL model are projected onto a two-dimensional pixel plane, and then the Euclidean distance between the projection coordinates of the three-dimensional joint and the coordinates of the positioned two-dimensional joint points is minimized by optimizing an energy function, so that the SMPL model can be fitted to a human body in a corresponding RGB image, and the motion postures of the human body are matched with each other. The method makes full use of three-dimensional space information contained in two-dimensional joint points, and proposes to establish a capsule collision body to solve the problem of limb cross penetration of an SMPL model in the gesture modeling process.
The energy function used by the SMPLIFY algorithm is as follows:
E(β,θ)＝EJ(β,θ；K,Jest)+λθEθ(θ)+λαEα(θ)+λspEsp(θ；β)+λβEβ(β) (1)
wherein, β and θ are body type parameters and posture parameters used by the SMPL model, and K is an internal reference matrix of the used camera, and mainly includes information such as focal length, distortion correction coefficient, resolution and the like of the camera. J. the design is a squareestFor the predicted two-dimensional human body joint point coordinates, λθ,λα,λsp,λβIs the weight of the respective energy function.
Distance constraint term based on joint points: penalizing the weighted 2D distance between the estimated joint point and the corresponding projected SMPL joint point, and the formula is as follows:
EJ(β,θ；K,Jest)＝∑jointωiρ(∏K(Rθ(J(β)i))-Jest,i) (2)
wherein, J (beta)iRepresenting the estimation of a function of the bone joint using shape parameters,. piKFor orthogonal projection of a parametric K-camera from three to two dimensions, Jest,iAnd ωiCoordinates and confidence degrees of the ith joint point obtained by a two-dimensional human body joint point positioning method are processed by using a Geman-McClure function rhoθRepresenting a global rigid transformation function related to the pose parameters. The specific process of the method is as follows: and after estimating camera parameters K, projecting the three-dimensional joint coordinates of the SMPL model to a two-dimensional image, calculating the weighted Euclidean distance between the predicted two-dimensional joint point coordinates and the corresponding two-dimensional projection coordinates of the three-dimensional joint, and taking the sum of the Euclidean distances of all joints as the energy value of the energy item.
The attitude constraint term: the elbows and knees which are not naturally bent are punished.
Wherein exp (theta)i) Mainly used for restricting the rotation angle of the joint, each joint of the SMPL model has three degrees of freedom, but the real human body joint can not rotate freely. Therefore, when the attitude parameter theta is solved, the parameter needs to be constrained within a certain range, and when the rotation angle of the joint exceeds a certain range, the energy value of the term is increased, so that the rotation angle of each joint of the SMPL model is constrained, and unreasonable attitude is avoided. Exponential functions severely penalize rotations of the elbow and knee that violate natural constraints.
Solving the constrained energy item of the attitude:
wherein, muθ,jExpressed is the mean, Σ, of the jth gaussian functionθ,jThe variance of the jth Gaussian function is shown, and when the three-dimensional posture of the human body is solved from the two-dimensional joint points, the condition that the solving result is not unique due to the lack of depth information can occur. In the SMPLIFY algorithm, near 100 in the CMU database is utilized firstTen thousand different body posture data generate a series of SMPL models, then a plurality of Gaussian distributions (namely Gaussian mixture models) are built by using the models, and the Gaussian mixture models are used for constraining the bias in the posture parameter theta solving process, so that the parameter solving space is further reduced, and more accurate results are searched in a plurality of possible body postures. In the above formula, gjThe weight of the jth of the 8 gaussian distributions is represented and N (-) represents the probability density function. Since the computation cost of calculating the negative logarithm of the weighted probability sum of all gaussian distributions is large, the algorithm uses the largest weighted probability in the gaussian distributions for approximation, and uses the constant c to reduce the error caused by the approximation.
Collision energy term:
the function of the energy item is mainly to restrict the posture and the body type of the SMPL model and avoid the phenomenon that different blocks in the three-dimensional human body model are fused with each other. In the posture modeling process based on the SMPL model, the accuracy of a three-dimensional skeleton built in the model is considered, and the problem that all parts of the body of the model are fused and penetrated with each other needs to be solved. To this end, the SMPLIFY algorithm approximates the SMPL model surface as a set of "capsules" to prevent collision and penetration between different regions of the human body. In the process of adjusting the posture and the body type of the three-dimensional human body model, whether limbs of the model cross and penetrate is calculated by utilizing a capsule collision body, and the energy value is large when capsule bodies collide with each other. In the above formula, C (θ, β) represents the center coordinates of the capsule collision volume, r (β) represents the radius of the capsule collision volume, σ (β) is equal to r (β)/3, and i (i) represents a set of spheres incompatible with the i-th sphere.
Body type constraint term:
the upper typeIn (1),
and the diagonal matrix obtained by principal component analysis in the process of training by using the SMPL model is shown. The main calculation of the item is the difference of body types between the current model and the SMPL template model, and the larger the difference of body type parameters beta of the current model and the SMPL template model is, the larger the energy value is, so that the body type of the model is restrained, and extreme conditions such as too small or high fatness of the obtained model can not occur in the process of solving the body type parameters.
The idea of HMR is as follows: the Human body Mesh reconstruction (HMR) method is a method for restoring a three-dimensional Human body model from end to end by using an RGB image and is realized based on a convolutional neural network. The HMR method uses the idea of competing with the generation of a network, and the entire network model can be represented as an encoder and an arbiter. The HMR method differs from the conventional two-stage method in that 2D human joint points are first inferred from two-dimensional images, and then three-dimensional models are predicted or three-dimensional parameters are calculated from the 2D joint points.
The HMR model network uses a GAN network structure and is divided into an encoder and a discriminator.
The loss function of the encoder is L3DAnd LreprojThe following formula is shown below. The regression of the SMPL model parameters was achieved by training the entire network with the following two loss functions.
L3D＝L3Djoints+L3DSMPL (7)
In the above formula, viFor confidence in the two-dimensional joint point, Π represents the orthogonal projection. L is3DSMPLRepresenting the loss function when trained using a three-dimensional dataset based on the SMPL model. Wherein
The labeled SMPL model parameters used by the training set are represented.
The loss function of the discriminator is LadvThe following formula is shown below. And learning the limit of the rotation angle of each human joint by using the following loss function to judge whether the model posture is reasonable.
minLadv(E)＝∑iEΘ～pe[(Di(E(I))-1)2] (12)
minL(Di)＝EΘ～pdata[(Di(Θ)-1)2]+EΘ～pe[(Di(E(I))2] (13)
Wherein E represents an encoder, and D represents a discriminator; in order to solve the problem of lack of a three-dimensional attitude data set, the HMR algorithm trains the network by using massive 2D unlabeled data and partial 3D labeled data. The loss function of the entire network training is shown as follows:
L＝λ(Lreproj+IL3D)+Ladv (14)
where λ is the weight of the encoder and the discriminator, and I is an exponential function, and when I is 1, it means that the network is trained using 3D labeled data.
The training process of the entire network can be briefly expressed as: firstly, input the ith picture phi of the training seti，ΦiOutputting image characteristics after being processed by an encoder (convolutional layer, VGG-19), performing parameter regression (full connected layer) on the output image characteristics, and outputting to obtain a vector theta with 85 dimensionsiWhere, the parameters s include zoom, rotation, and translation parameters of the cameraR, t and SMPL model body type parameter beta and attitude parameter theta. Parameters to be predicted
And
inputting the three-dimensional joint coordinates into the SMPL model, and obtaining the three-dimensional joint coordinates of the model according to the built-in function J of the SMPL model
And then projecting the three-dimensional joint of the model to a two-dimensional image through a formula (11) by combining with camera parameters to obtain a predicted two-dimensional joint position
From 2D joint data in the training set used
And 3D annotation data
And (3) completing the loss calculation of the encoder layer through two loss functions (7) and (10) of the encoder layer, and finally performing back propagation of errors by combining with a loss function (12) of a discriminator to adjust the weight of each node of the network. And (5) circulating the process for a plurality of times until the whole network is trained after convergence.
To avoid unreasonable poses of the human body model generated by the encoder, the HMR algorithm uses a discriminator to constrain the network. Due to the parameter separable characteristic of the SMPL model, the HMR algorithm trains discriminators for body type parameters and posture parameters respectively and independently, and additionally trains a discriminator for restricting the overall posture of the model and a discriminator for restricting the overall body type of the model, so that 26 discriminators are trained. The output range of the discriminator is [0,1 ]]Confidence of (D)iTo represent the accuracy of the data and to determine its stability in a least squares formulation. Let E (I) denote the discriminator, the penalty function used by the encoder can be expressed as shown in equation (13)The loss function for a single discriminator is calculated using equation (12).
After obtaining the three-dimensional human body model, performing a meridian-acupoint positioning process by using a "bone degree" deflection method, first obtaining body surface joint points on the three-dimensional human body model, such as easily directly obtained body surface joint points, e.g., knee joint points, elbow joint points, etc., and then obtaining bone degree inches based on the obtained body surface joint points, for example: the shoulder blade inner margin → posterior median line is 3 inches, the shoulder peak margin → posterior median line is 8 inches, and then the specific meridian point location is carried out based on the initial reference point and the bone degree cun selected in advance, it should be noted that when the initial reference point is selected, intuitive and easily located meridian point positions are often selected, for example, an elbow joint point, a wrist joint point, a knee joint point, two joint points of a thumb and the like; in the process, the initial reference point is equivalent to the initial coordinate, and the bone length dimension is equivalent to providing a distance measuring basis between the meridians and the points.
According to the channel and acupoint positioning method provided by the invention, the three-dimensional human body model is obtained based on a single RGB image by utilizing an SMPLIFY method or an HMR method, the obtained three-dimensional human body model is used as a channel and acupoint positioning basis, then the pre-selected initial reference point is used as a reference position, and the bone length dimension is used as a distance measuring basis between channels and acupoints to carry out specific channel and acupoint positioning, so that the channel and acupoint positioning efficiency and the channel and acupoint positioning precision are improved.
According to the method for locating the acupuncture points provided by the invention, the acquiring of the joint points in the target human body image and the construction of the three-dimensional human body model of the target human body based on the joint points specifically comprise the following steps: and acquiring joint points in the target human body image by using an Open Pose method or an Alpha Pose method, and acquiring a three-dimensional human body model of the target human body based on the joint points.
It should be noted that the acquired human body image is input into an Open pos or Alpha pos model, then joint coordinates of the human body and a confidence thereof are output, and then the obtained joint coordinates are used to construct a three-dimensional human body model.
According to the channel and point positioning method provided by the invention, the three-dimensional human body model is obtained through the SMPL model based on a single RGB image, and the next specific channel and point positioning process is carried out based on the three-dimensional human body model, so that the channel and point positioning efficiency and positioning accuracy are improved.
According to the acupoint positioning method provided by the invention, the acupoint positioning is carried out based on the initial reference point and the bone size, and the method specifically comprises the following steps:
acquiring the three-dimensional space coordinates of the initial reference points, and taking the initial reference points as reference points;
and taking the bone length dimension as a measurement basis of the distance between the channels and the points to obtain the three-dimensional space coordinate of the target channels and points.
The method comprises the steps of giving a three-dimensional space coordinate of an initial reference point, taking the three-dimensional space coordinate of the initial reference point as a reference coordinate, taking the obtained bone degree size as a measurement basis of the distance between acupuncture points, determining the positions of other acupuncture points, and feeding back in a three-dimensional space coordinate mode; the initial reference points are not fixed, and the initial reference points selected when the acupuncture points of different parts of the human body are positioned can be changed, so long as the positions have clear physical meanings and are convenient to position the acupuncture points by using a bone degree fracture quantity positioning method, the positions can be used as the initial reference points. For example, the following steps are carried out: joint points at the centers of the shoulders and the back can be used for calculating the folding component (8 inches) of the shoulder peak edge → the posterior median line, and are used for determining the transverse distance of meridian points at the shoulders and the back, and then the acupuncture points at the back and the waist are positioned by taking the transverse distance as a standard; ② joint points of wrist and elbow, the fold length inch (12 inch) of elbow transverse striation (elbow tip) → palmar (dorsal) transverse striation can be calculated for determining the longitudinal distance of forearm meridian point, then the acupoint of upper limb can be positioned according to the above-mentioned rule.
According to the channel and point positioning method provided by the invention, the three-dimensional space coordinate of the initial reference point is determined, and the three-dimensional space coordinate is used as a reference coordinate to perform coordinate positioning on other channel and point positions, so that the accuracy of channel and point positioning is improved, and the visualization of the channel and point positioning process is further improved by the coordinate display of the channel and point positions.
According to the acupoint positioning method provided by the invention, the acupoint positioning is carried out based on the initial reference point and the bone size, and the method specifically comprises the following steps:
and acquiring the inter-digital width of the thumb in the three-dimensional human body model, correcting the bone degree cun based on the inter-digital width, and performing acupoint location based on the initial reference point and the corrected bone degree cun.
It should be noted that, the inter-digital joint width of the thumb in the three-dimensional human body model is acquired, the inter-digital joint width is defined as 1 inch, the inter-digital joint width is used for correcting and comparing the acquired other bone dimension, and the initial reference point and the corrected bone dimension are used for performing acupoint positioning.
According to the meridian point positioning method provided by the invention, the width of the interphalangeal joint of the thumb in the obtained three-dimensional human body model is taken as 1 inch, and then the obtained bone degree size is taken as a basis for correcting the obtained bone degree size, so that the obtained bone degree size is more targeted, and the accuracy of meridian point positioning is improved.
According to the present invention, there is provided a method for locating a meridian point, wherein the method further comprises:
when a second positioning method is selected, inputting the three-dimensional human body model into a trained three-dimensional human body positioning model to obtain the three-dimensional human body model for completing acupoint positioning; the three-dimensional human body positioning model is trained by utilizing at least one three-dimensional human body model for completing acupoint labeling.
It should be noted that, on the basis of obtaining the three-dimensional human body model, the neural network model is used to realize the channel and point positioning process, and the complete process is as follows:
constructing a three-dimensional human body positioning model for realizing the channel and acupoint positioning process of the three-dimensional human body model; training the three-dimensional human body positioning model by utilizing a training set to obtain a trained three-dimensional human body positioning model, wherein the training set comprises a plurality of three-dimensional human body models which are used for completing acupoint positioning marking; inputting the three-dimensional human body model to be subjected to channel and acupoint positioning into the three-dimensional human body positioning model, and obtaining the three-dimensional human body model subjected to channel and acupoint positioning through the positioning and labeling process of the three-dimensional human body positioning model.
According to the channel and acupoint positioning method provided by the invention, on the basis of obtaining the three-dimensional human body model, the three-dimensional human body positioning neural network model is constructed, and the three-dimensional human body model to be positioned is input into the three-dimensional human body positioning model to obtain the three-dimensional human body model for completing channel and acupoint positioning, so that the channel and acupoint positioning precision and the channel and acupoint positioning efficiency are improved based on the process.
FIG. 2 is a schematic flow chart of a three-dimensional human body model modeling process provided by the present invention, as shown in FIG. 2, a human body is segmented from a background by using a human body segmentation technique; the three-dimensional human body modeling can be guided by taking the segmentation result as reference, so that the modeling is more accurate. Instead of using the segmentation map, the three-dimensional human body modeling can be performed by directly using the acquired image including the human body by using the SMPLIFY algorithm or the HMR method.
Fig. 3 is a schematic view of the bone degree folding dimension of the back and waist portion provided by the present invention, as shown in fig. 3, the folding dimension (8 inches) of the shoulder edge → posterior median line is calculated based on the joint points of the shoulder and the back center obtained by the human body joint point positioning method, and is used to determine the transverse distance of the meridian points of the shoulder and the back, and then the acupuncture points of the back and waist portion are positioned based on the calculated transverse distance.
Fig. 4 is a schematic diagram of the location of lung meridian points of hand taiyin provided by the present invention, and as shown in fig. 4, the location process of lung meridian points of hand taiyin is as follows:
step1, back center joint point O using OpenPose positioning1And left shoulder joint point O2As an initial reference point, the folding point of "acromion → posterior midline" (8 cun), i.e. the line connecting the two joints of the center of the back and the left shoulder, is calculated using fig. 3
The length of the foot is 8 inches, so that the length corresponding to 1 inch can be calculated;
step2, back center point O1As a reference point, along
Moving 6 cun to the left to obtain Yunmen acupoint;
step3, taking the obtained Yunmen acupoint as a reference point, and comparing the points with the reference point
Moving downwards in the vertical direction for 1 inch to obtain "Zhongfu" acupoints;
step4, similarly, calculating the positions of the rest acupuncture points by means of the positioned acupuncture points according to the position relation of each acupuncture point in the hand taiyin lung meridian point national standard position table in the national standard of the people's republic of China, acupuncture point position;
it should be noted that, since the above process is performed on a three-dimensional human body model, three-dimensional space coordinates of all joint points can be calculated in a world coordinate system by using each joint point obtained by OpenPose positioning, and then the three-dimensional space coordinates of each acupoint are calculated according to the mapping of the corresponding relationship.
Fig. 5 is a schematic structural view of the acupoint positioning device provided by the present invention, and as shown in fig. 5, the acupoint positioning device includes: an obtaining module 510, a processing module 520, a constructing module 530, and a positioning module 540, wherein:
an obtaining module 510 configured to: obtaining a single RGB image;
a processing module 520 configured to: preprocessing the single RGB image to obtain a target human body image;
a build module 530 to: acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points;
a positioning module 540, configured to: and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
According to the acupoint positioning device, the three-dimensional human body model is obtained based on a single RGB image by utilizing a Smplify three-dimensional human body reconstruction method, the obtained three-dimensional human body model is used as a acupoint positioning basis, then the pre-selected initial reference point is used as a reference position, and the bone length dimension is used as a distance measuring basis between acupoints, so that specific acupoint positioning is carried out, and the acupoint positioning efficiency and the acupoint positioning accuracy are improved.
According to the acupoint positioning device provided by the present invention, the processing module 520, when being configured to pre-process the single RGB image to obtain the target human body image, is specifically configured to: and carrying out image enhancement processing and/or human body segmentation processing on the single RGB image to obtain the target human body image.
According to the acupoint positioning device provided by the invention, the image enhancement processing is carried out on a single RGB image, the imaging quality of the image is improved, the specific positions of all parts of the human body in the image are determined for the human body segmentation processing of the image, the matching degree of the obtained three-dimensional human body model and the actual human body shape is improved based on the above processes, and the accuracy of acupoint positioning is finally improved.
According to the acupoint positioning device provided by the present invention, the constructing module 530, when being configured to acquire the joint points in the target human body image, is specifically configured to, when constructing the three-dimensional human body model of the target human body based on the joint points: acquiring joint points in the target human body image by using an Open Pose method or an Alpha Pose method, acquiring skeleton coordinates of the target human body based on the joint points, and acquiring a three-dimensional human body model of the target human body based on the skeleton coordinates.
According to the acupoint positioning device provided by the invention, the three-dimensional human body model is obtained through the SMPL model based on a single RGB image, and the following specific acupoint positioning process is carried out based on the three-dimensional human body model, so that the acupoint positioning efficiency and the positioning precision are improved.
According to the acupoint locating device provided by the invention, the locating module 540 is specifically used for: acquiring the three-dimensional space coordinates of the initial reference points, and taking the initial reference points as reference points; and taking the bone length dimension as a measurement basis of the distance between the channels and the points to obtain the three-dimensional space coordinate of the target channels and points.
According to the meridian point positioning device provided by the invention, the three-dimensional space coordinate of the initial reference point is determined, and the three-dimensional space coordinate is used as a reference coordinate to perform coordinate positioning on other meridian point positions, so that the accuracy of meridian point positioning is improved, and meanwhile, the visualization of the meridian point positioning process is further improved due to the coordinate display of the meridian point positions.
According to the acupoint locating device provided by the invention, the locating module 540 is specifically used for: and acquiring the inter-digital width of the thumb in the three-dimensional human body model, correcting the bone degree cun based on the inter-digital width, and performing acupoint location based on the initial reference point and the corrected bone degree cun.
According to the meridian point positioning device provided by the invention, the width of the interphalangeal joint of the thumb in the obtained three-dimensional human body model is taken as 1 inch, and then the obtained bone degree size is taken as a basis for correcting the obtained bone degree size, so that the obtained bone degree size is more targeted, and the accuracy of meridian point positioning is improved.
According to the present invention, the apparatus 500 is further configured to: when a second positioning method is selected, inputting the three-dimensional human body model into a trained three-dimensional human body positioning model to obtain the three-dimensional human body model for completing acupoint positioning; the three-dimensional human body positioning model is trained by utilizing at least one three-dimensional human body model for completing acupoint labeling.
According to the meridian point positioning device, on the basis of obtaining the three-dimensional human body model, the three-dimensional human body positioning neural network model is constructed, the three-dimensional human body model to be positioned is input into the three-dimensional human body positioning model, the three-dimensional human body model for completing meridian point positioning is obtained, and the accuracy of meridian point positioning and the efficiency of meridian point positioning are improved based on the process.
Fig. 6 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 6: a processor (processor)610, a communication Interface (Communications Interface)620, a memory (memory)630 and a communication bus 640, wherein the processor 610, the communication Interface 620 and the memory 630 communicate with each other via the communication bus 640. The processor 610 may invoke logic instructions in the memory 630 to perform a method of acupoint positioning, the method comprising: obtaining a single RGB image; preprocessing the single RGB image to obtain a target human body image; acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points; and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
In addition, the logic instructions in the memory 630 may be implemented in software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the method of acupoint positioning provided by the above methods, the method comprising: obtaining a single RGB image; preprocessing the single RGB image to obtain a target human body image; acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points; and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform the above-provided acupoint locating methods, the method comprising: obtaining a single RGB image; preprocessing the single RGB image to obtain a target human body image; acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points; and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (9)
1. A method of locating a point, comprising:
obtaining a single RGB image;
preprocessing the single RGB image to obtain a target human body image;
acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points;
and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
2. The method for locating acupuncture points according to claim 1, wherein the preprocessing the single RGB image to obtain the target human body image specifically includes:
and carrying out image enhancement processing and/or human body segmentation processing on the single RGB image to obtain the target human body image.
3. The method according to claim 1, wherein the acquiring of the joint points in the target human body image and the constructing of the three-dimensional human body model of the target human body based on the joint points comprise:
and acquiring joint points in the target human body image by using an Open Pose method or an Alpha Pose method, and acquiring a three-dimensional human body model of the target human body based on the joint points.
4. The method of claim 1, wherein the performing the acupoint positioning based on the initial reference point and the bone size dimension comprises:
and acquiring the three-dimensional space coordinates of the initial reference points, taking the initial reference points as reference points, and taking the bone degree dimensions as measurement basis of the distance between the channels and the points to obtain the three-dimensional space coordinates of the target channels and points.
5. The method of claim 1, wherein the performing the acupoint positioning based on the initial reference point and the bone size dimension comprises:
and acquiring the inter-digital width of the thumb in the three-dimensional human body model, correcting the bone degree cun based on the inter-digital width, and performing acupoint location based on the initial reference point and the corrected bone degree cun.
6. The method of claim 1, further comprising:
when a second positioning method is selected, inputting the three-dimensional human body model into a trained three-dimensional human body positioning model to obtain the three-dimensional human body model for completing acupoint positioning; the three-dimensional human body positioning model is trained by utilizing at least one three-dimensional human body model for completing acupoint labeling.
7. A acupoint positioning device, comprising:
an acquisition module to: obtaining a single RGB image;
a processing module to: preprocessing the single RGB image to obtain a target human body image;
a build module to: acquiring joint points in the target human body image, and constructing a three-dimensional human body model of the target human body based on the joint points;
a positioning module to: and when the first positioning method is selected, obtaining body surface bone nodes of the three-dimensional human body model, obtaining bone degree cunnings based on the body surface bone nodes, and performing acupoint positioning based on the initial reference points and the bone degree cunnings.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the steps of the method of acupuncture point location according to any one of claims 1 to 6 are implemented when the program is executed by the processor.
9. A non-transitory computer readable storage medium, having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the acupoint positioning method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111187653.1A CN114092634A (en) | 2021-10-12 | 2021-10-12 | Channel and acupoint positioning method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111187653.1A CN114092634A (en) | 2021-10-12 | 2021-10-12 | Channel and acupoint positioning method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114092634A true CN114092634A (en) | 2022-02-25 |
Family
ID=80296783
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111187653.1A Pending CN114092634A (en) | 2021-10-12 | 2021-10-12 | Channel and acupoint positioning method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114092634A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI807678B (en) * | 2022-03-11 | 2023-07-01 | 亞東學校財團法人亞東科技大學 | Interactive massage part generating method and system |
| CN116531248A (en) * | 2023-05-10 | 2023-08-04 | 上海芯兀极智能科技有限公司 | Human body acupoint positioning method based on multi-sensing fusion |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102743282A (en) * | 2012-07-03 | 2012-10-24 | 成都中医药大学 | Intelligent acupuncture point locating device and method based on image identification |
| CN110930374A (en) * | 2019-11-13 | 2020-03-27 | 北京邮电大学 | Acupoint positioning method based on double-depth camera |
| US20200126297A1 (en) * | 2018-10-17 | 2020-04-23 | Midea Group Co., Ltd. | System and method for generating acupuncture points on reconstructed 3d human body model for physical therapy |
| CN112184705A (en) * | 2020-10-28 | 2021-01-05 | 成都智数医联科技有限公司 | Human body acupuncture point identification, positioning and application system based on computer vision technology |
-
2021
- 2021-10-12 CN CN202111187653.1A patent/CN114092634A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102743282A (en) * | 2012-07-03 | 2012-10-24 | 成都中医药大学 | Intelligent acupuncture point locating device and method based on image identification |
| US20200126297A1 (en) * | 2018-10-17 | 2020-04-23 | Midea Group Co., Ltd. | System and method for generating acupuncture points on reconstructed 3d human body model for physical therapy |
| CN110930374A (en) * | 2019-11-13 | 2020-03-27 | 北京邮电大学 | Acupoint positioning method based on double-depth camera |
| CN112184705A (en) * | 2020-10-28 | 2021-01-05 | 成都智数医联科技有限公司 | Human body acupuncture point identification, positioning and application system based on computer vision technology |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI807678B (en) * | 2022-03-11 | 2023-07-01 | 亞東學校財團法人亞東科技大學 | Interactive massage part generating method and system |
| CN116531248A (en) * | 2023-05-10 | 2023-08-04 | 上海芯兀极智能科技有限公司 | Human body acupoint positioning method based on multi-sensing fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Pishchulin et al. | Building statistical shape spaces for 3d human modeling | |
| US11127163B2 (en) | Skinned multi-infant linear body model | |
| JP7061694B2 (en) | Image processing methods and equipment, imaging equipment, and storage media | |
| Hesse et al. | Learning and tracking the 3D body shape of freely moving infants from RGB-D sequences | |
| Kadleček et al. | Reconstructing personalized anatomical models for physics-based body animation | |
| CN114092634A (en) | Channel and acupoint positioning method and device, electronic equipment and storage medium | |
| CN109636910A (en) | A kind of cranium face restored method generating confrontation network based on depth | |
| Claes | A robust statistical surface registration framework using implicit function representations-application in craniofacial reconstruction | |
| WO2020147791A1 (en) | Image processing method and device, image apparatus, and storage medium | |
| Lee et al. | Im2hands: Learning attentive implicit representation of interacting two-hand shapes | |
| Zhang et al. | 3D statistical head modeling for face/head-related product design: a state-of-the-art review | |
| Chi et al. | A new parametric 3D human body modeling approach by using key position labeling and body parts segmentation | |
| WO2020147797A1 (en) | Image processing method and apparatus, image device, and storage medium | |
| Komaritzan et al. | Inside humans: Creating a simple layered anatomical model from human surface scans | |
| CN116721471A (en) | Multi-person three-dimensional attitude estimation method based on multi-view angles | |
| CN117115922A (en) | Seat body forward-bending evaluation method, system, electronic equipment and storage medium | |
| Bauer et al. | Anatomical mirroring: Real-time user-specific anatomy in motion using a commodity depth camera | |
| CN116433812A (en) | Method and device for generating virtual character by using 2D face picture | |
| US20220198696A1 (en) | System for determining body measurement from images | |
| CN115954105A (en) | Whole body anatomical structure personalized modeling and posture deformation algorithm for human body biomechanical simulation | |
| CN113902845A (en) | Motion video generation method and device, electronic equipment and readable storage medium | |
| WO2020147794A1 (en) | Image processing method and apparatus, image device and storage medium | |
| JP2023090566A (en) | Program, apparatus and method for statistically analyzing flesh-based body shape from skin model | |
| Zientara et al. | Individualised avatars with complete anatomy constructed from the ANSUR II 3-D anthropometric database | |
| CN112185508A (en) | Eye exercises quality evaluation system based on artificial intelligence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220225 |
|
| RJ01 | Rejection of invention patent application after publication |