A soft tissue deformation simulation method and device based on a cellular neural network

Abstract

The application belongs to the technical field of soft tissue deformation simulation, and particularly relates to a soft tissue deformation simulation method and device based on a cell neural network. The method comprises the following steps: constructing a three-dimensional cell neural network model; applying an external force to the three-dimensional cell neural network model, and calculating a cell current of a cell in the three-dimensional cell neural network model according to the applied external force; calculating a state voltage of the cell according to the cell current; calculating a displacement of the cell according to the state voltage; and performing deformation simulation on the three-dimensional cell neural network model according to the displacement of the cell. The application has good generalization, and can realize fast modeling while maintaining simulation real-time performance and simulation effect.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of soft tissue deformation simulation, and particularly relates to a soft tissue deformation simulation method and device based on a cellular neural network. BACKGROUND

[0002] Non-contact medical technology provides protection for the current medical industry due to its ability to greatly avoid cross-infection between doctors and patients. In the field of non-contact medical technology, human-computer interaction has always been a research focus. Especially in the simulation of soft tissue deformation of the human body, it is crucial for medical staff to simulate soft tissue deformation with high real-time performance and realistic effects. In recent years, with the development of computer technology, neural network technology has shown excellent performance in various fields. This technology has also attracted attention in soft tissue deformation modeling in surgical simulation. It provides a new direction for the study of soft tissue deformation modeling with its parallel processing and ability to fully approximate any complex nonlinear relationship.

[0003] Currently, neural network technology in soft tissue deformation modeling mostly adopts the method of obtaining a neural network proxy model through training. Although the trained model can achieve good real-time performance during simulation, the difficulty in collecting training samples and the long training time make the existing model still only select a small number of parameters for training. This makes the model have poor generalization. SUMMARY

[0004] To solve the problems of the prior art, the application provides a soft tissue deformation simulation method and device based on a cellular neural network, which has good generalization and can achieve fast modeling while maintaining simulation real-time performance and simulation effects.

[0005] To solve the problems of the prior art, the technical solution provided by the application is as follows:

[0006] A soft tissue deformation simulation method based on a cellular neural network, comprising,

[0007] constructing a three-dimensional cellular neural network model;

[0008] applying an external force to the three-dimensional cellular neural network model, and calculating the cell current of the cells in the three-dimensional cellular neural network model according to the applied external force;

[0009] calculating the state voltage of the cells according to the cell current of the cells;

[0010] calculating the displacement of the cells according to the state voltage of the cells;

[0011] performing deformation simulation on the three-dimensional cellular neural network model according to the displacement of the cells.

[0012] Preferably, the cell current of the cell in the three-dimensional cell neural network model is calculated according to the applied external force, comprising,

[0013] ||I ijk (t)|| = ||σ ijk (t)|

[0014] I ijk (t) is defined as the cell current of the cell C(i,j,k) at the coordinate (i,j,k) in the three-dimensional cell neural network model at time t, and ||I ijk (t)|| is the numerical value of I ijk (t);

[0015] When the cell C(i,j,k) is a force point, the cell current direction of the cell C(i,j,k) at time t is defined as consistent with the direction of the external force applied to the cell C(i,j,k) at time t;||σ ijk (t)|| is the numerical value of the stress σ ijk (t) of the cell C(i,j,k) at time t;

[0016] When the cell C(i,j,k) is a non-force point, I ijk (t) is 0.

[0017] Preferably, the state voltage of the cell is calculated according to the cell current of the cell, comprising,

[0018]

[0019] Wherein, v xijk (t) is the state voltage of the cell C(i,j,k) at the coordinate (i,j,k) in the three-dimensional cell neural network model at time t, R x is the cell resistance, I ijk (t) is the cell current of the cell C(i,j,k) at time t, C(a,b,c) is the adjacent cell of the cell C(i,j,k), N r (i,j,k) is the set of adjacent cells of the cell C(i,j,k), A (i，j，k；a，b，c) (t) is the feedback coefficient of the interaction between the cell C(i,j,k) and the adjacent cell C(a,b,c) at time t, and v yabc (t) is the output voltage of the adjacent cell C(a,b,c) at time t.

[0020] Preferably, the calculation formula of A (i，j，k；a，b，c) (t) is as follows:

[0021]

[0022] Wherein, α is the energy transmission coefficient, E ijk and Eabc The elastic moduli of cell C(i,j,k) and neighboring cell C(a,b,c), respectively, and ||E ijk +E abc || represents E ijk +E abc The value of cos <v yabc (t), L (i，j，k；a，b，c) (t)> represents the output voltage v of neighboring cell C(a, b, c) at time t. yabc The cosine of the angle between the direction of (t) and the vector direction originating from cell C(i,j,k) and ending at neighboring cell C(a,b,c) at time t.

[0023] Preferably, it also includes the R x The calculation formula is as follows:

[0024]

[0025] Among them, ||R x ||For R x The numerical value of ||l|| refers to the initial interval length l between adjacent cells in the three-dimensional cellular neural network model; ||A|| is the numerical value of the cross-sectional area A of the cell; β is the cell displacement coefficient; ||E ijk || represents the elastic modulus E of cell C(i,j,k). ijk The value.

[0026] Preferably, the v yabc The formula for calculating (t) is as follows:

[0027]

[0028] Among them, v xabc (t) represents the state voltage of the neighboring cell C(a, b, c) at time t. yabc Direction and v xabc The directions of (t) are the same, and K is the saturation voltage of the cell.

[0029] Preferably, determining the cell displacement based on the cell's state voltage includes,

[0030] ||Δl ijk (t)||=β||v xijk (t)||||E ijk ||

[0031] Where, Δl is defined ijk (t) represents the displacement of cell C(i,j,k) at coordinate (i,j,k) in the 3D cellular neural network model at time t, ||Δl ijk (t)|| is Δli ja value of k(t); v xijk a state voltage of the cell C(i,j,k) at time t; ||v xijk a value of v xijk a value of Δl ijk a direction of v xijk a direction of v ijk a cell displacement coefficient; E ijk a modulus of elasticity of the cell C(i,j,k); ||E ijk a value of E

[0032] A soft tissue deformation simulation device based on a cellular neural network, comprising:

[0033] a construction module configured to construct a three-dimensional cellular neural network model;

[0034] an external force application module configured to apply an external force to the three-dimensional cellular neural network model;

[0035] a cell current calculation module configured to calculate a cell current of a cell in the three-dimensional cellular neural network model according to the applied external force;

[0036] a state voltage calculation module configured to calculate a state voltage of the cell according to the cell current of the cell;

[0037] a displacement calculation module configured to calculate a displacement of the cell according to the state voltage of the cell;

[0038] a deformation simulation module configured to perform deformation simulation on the three-dimensional cellular neural network model according to the displacement of the cell.

[0039] Preferably, the state voltage calculation module is specifically configured to calculate the state voltage of the cell according to the following formula:

[0040]

[0041] wherein v xijk is a state voltage of a cell C(i,j,k) at a coordinate (i,j,k) in the three-dimensional cellular neural network model at time t, R x is a cell resistance, I ijk is a cell current of the cell C(i,j,k) at time t, C(a,b,c) is an adjacent cell of the cell C(i,j,k), N r (i,j,k) is a set of adjacent cells of the cell C(i,j,k), A (i，j,k；a，b，c) is a feedback coefficient of the cell C(i,j,k) and the adjacent cell C(a,b,c) at time t, v yabc is an output voltage of the adjacent cell C(a,b,c) at time t.

[0042] Preferably, the displacement obtaining module is specifically used for obtaining the displacement of the cell according to the following formula:

[0043] ||Δl ijk (t)||=β||v xijk (t)||||E ijk ||

[0044] wherein, Δl ijk (t) is the displacement of the cell C(i,j,k) at the coordinate (i,j,k) in the three-dimensional cell neural network model at the time t, ||Δl ijk (t)|| is the numerical value of Δl j k(t); v xijk (t) is the state voltage of the cell C(i,j,k) at the time t, ||v xijk (t)|| is the numerical value of v xijk (t); the direction of Δl ijk (t) is the same as the direction of v xijk (t); β is a cell displacement coefficient; E ijk is the elastic modulus of the cell C(i,j,k), ||E ijk || is the numerical value of E ijk .

[0045] Advantages of the present application:

[0046] The present application directly utilizes the characteristics of local interconnection and parallel computation of the cell neural network to construct a cell deformation model; converts the external force into a cell current in the cell neural network, so as to realize the conversion between the kinetic energy of soft tissue and the bioelectric energy in the cell neural network; utilizes the local interconnection of the cells in the cell neural network, and realizes the transmission of the kinetic energy between soft tissues by redefining the energy transmission rules between the cells; utilizes the finite element method to calculate the cell parameters in the cell neural network, so as to realize the adaptive parameter selection between heterogeneous soft tissues, and improve the generalization of the model; the model realizes fast modeling while maintaining the simulation real-time, and has a realistic deformation effect.

[0047] The present application utilizes the high real-time provided by the parallel computation of the neural network, avoids the problems of poor generalization and long construction time of the current trained model, and scientifically determines the parameters of the cell neural network by the finite element method, so as to realize the realistic deformation effect of the model. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 The present application provides a schematic diagram of the soft tissue deformation simulation method based on the cell neural network. DETAILED DESCRIPTION

[0049] The application will be further described in connection with the embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the application, and cannot be used to limit the protection scope of the application.

[0050] Embodiment one

[0051] The embodiment of the application provides a soft tissue deformation simulation method based on a cell neural network, referring to Figure 1 , comprising,

[0052] S1, constructing a three-dimensional cell neural network model;

[0053] S2, applying an external force to the three-dimensional cell neural network model, and calculating a cell current of a cell in the three-dimensional cell neural network model according to the applied external force, so as to convert the external force applied by a virtual instrument into the cell current in the cell neural network;

[0054] ||I ijk (t)||=||σ ijk (t)||

[0055] I ijk (t) is defined as the cell current of a cell C(i, j, k) at a coordinate (i, j, k) in the three-dimensional cell neural network model at a time t, and ||I ijk (t)|| is the value of I ijk (t).

[0056] When the cell C(i, j, k) is a force point, the cell current direction of the cell C(i, j, k) at the time t is defined as being consistent with the direction of the external force suffered by the cell C(i, j, k) at the time t, and ||σ ijk (t)|| is the value of the stress σ ijk (t) of the cell C(i, j, k) at the time t.

[0057] When the cell C(i, j, k) is a non-force point, I ijk (t) is 0.

[0058] The application connects the applied external force and the cell current of the cell in the three-dimensional cell neural network model through the formula, converts the applied external force into the cell current in the three-dimensional cell neural network model, and simulates the cell deformation based on the three-dimensional cell neural network model.I ijk (t) is equal to the value of σ ijk (t). The value of σ ijk (t) is HN, and I ijk (t) is HA. The value in the following formula has the same meaning.

[0059] S3, calculating a state voltage of the cell according to the cell current, comprising,

[0060]

[0061] Among them, v xijk (t) represents the state voltage of cell C(i, j, k) at coordinate (i, j, k) in the three-dimensional cellular neural network model at time t, R. x Let N be the cell resistance, C(a, b, c) be the neighboring cells of cell C(i, j, k), and N be the cell resistance. r (i, j, k) is the set of neighboring cells of cell C(i, j, k), A (i，j，k；a，b，c) (t) is the feedback coefficient of the interaction between cell C(i,j,k) and neighboring cell C(a,b,c) at time t, v yabc (t) represents the output voltage of the adjacent cell C(a, b, c) at time t, v xijk The direction of (t) depends on v at time t. yabc (t) and I ijk The direction of (t) is calculated using the formula above.

[0062] This application redefines the rules of intercellular energy transfer based on mechanical principles, simplifies the cell state equation, and incorporates the neighboring cell C(a, b, c) into the cell state equation as the control coefficient B(i, j, k; a, b, c) that influences the state of cell C(i, j, k), as well as the state voltage change rate coefficient C of cell C(i, j, k). x Setting it to 0 yields this equation.

[0063] A (i，j，k；a，b，c) The formula for calculating (t) is as follows:

[0064]

[0065] Where α is a user-defined energy transfer coefficient, which can be adjusted according to the realism of the deformation simulation results; E ijk and E abc The elastic moduli of cell C(i,j,k) and neighboring cell C(a,b,c), respectively, and ||E ijk +E abc || represents E ijk +E abc The value of cos <v yabc (t), L (i，j，k；a，b，c) (t)> represents the output voltage v of neighboring cell C(a, b, c) at time t. yabc The cosine of the angle between the direction of (t) and the vector direction starting from cell C(i,j,k) and ending at neighboring cell C(a,b,c) at time t is used to simulate the influence of a cell on surrounding cells and to simulate the energy transfer process.

[0066] R xThe calculation formula is as follows:

[0067]

[0068] Wherein, ||R x is the value of R x ; ||l|| is the value of the initial interval length l between adjacent cells in the three-dimensional cell neural network model; ||A|| is the value of the cross-sectional area A of the cell; β is the cell displacement coefficient, which can be adjusted according to the actual degree of deformation simulation results; ||E ijk is the value of the elastic modulus E ijk of the cell C(i, j, k).

[0069] R x is determined by finite element simulation:

[0070] According to the finite element method, when an external force acts on the cross section of the finite element, the following can be obtained:

[0071]

[0072] Wherein, M is the unit mass of the finite element, that is, the mass of the stressed cell C(i, j, k), is the second derivative of the displacement X of the finite element with respect to time t, A is the cross-sectional area of the stressed cell C(i, j, k); E ijk represents the elastic modulus of the stressed cell C(i, j, k), ε ijk (t) is the strain of the stressed cell C(i, j, k) at time t; F ex (t) is the external force acting on the stressed cell C(i, j, k) at time t. When the cell is in a stable state, the second derivative of the displacement X of the stressed cell with respect to time t is 0.

[0073] The state voltage formula of the cell is transformed into the following form:

[0074]

[0075] The interaction of the set N r (i, j, k) of adjacent cells on the stressed cell C(i, j, k) and the current I ijk (t) of the stressed cell C(i, j, k) are collectively regarded as the external energy acting on the stressed cell C(i, j, k) equal to F ex (t), so I can be approximately obtained:

[0076] ||v xijk (t)|| = ||R x ||A||E ijk ||εijk (t)||

[0077] wherein, ||v xijk (t)|| is the value of v xijk (t), ||ε ijk (t)|| is the value of ε ijk (t) ;

[0078] wherein, ε ijk (t) reflects the displacement Δl ijk (t) of the stressed cell C(i, j, k), and the relationship between Δl ijk (t) and the initial interval length l between adjacent cells in the three-dimensional cell neural network model is:

[0079]

[0080] wherein, the displacement calculation formula in step S4 is used, ||Δl ijk (t)|| is the value of Δl yabc (t) ;

[0081] After the arrangement, we have:

[0082]

[0083] v xabc (t) is calculated according to the following formula:

[0084]

[0085] wherein, wherein, v yabc (t) is the state voltage of the adjacent cell C(a, b, c) at time t, the direction of v xabc (t) is the same as that of v xabc (t), K is the saturation voltage of the cell, which is set as 1. The calculation of v xijk (t) is the same as that of v

[0086] S4, calculating the displacement of the cell according to the state voltage of the cell:

[0087] ||Δl ijk (t)|| = β||v xijk (t)||||E ijk ||

[0088] wherein, Δl ijk (t) is defined as the displacement of the cell C(i, j, k) at coordinate (i, j, k) in the three-dimensional cell neural network model at time t, ||Δl ijk (t)|| is the value of Δl ijk (t) ; v xijk(t) is the state voltage of the cell C(i,j,k) at time t, ||v xijk (t) is the state voltage of the cell C(i,j,k) at time t, ||v xijk (t) is the state voltage of the cell C(i,j,k) at time t, ||v ijk (t) is the state voltage of the cell C(i,j,k) at time t, ||v xijk (t) is the state voltage of the cell C(i,j,k) at time t, ||v ijk (t) is the state voltage of the cell C(i,j,k) at time t, ||v ijk (t) is the state voltage of the cell C(i,j,k) at time t, ||v ijk (t) is the state voltage of the cell C(i,j,k) at time t, ||v

[0089] Embodiment Two

[0090] A soft tissue deformation simulation device based on a cellular neural network, comprising:

[0091] a construction module configured to construct a three-dimensional cellular neural network model;

[0092] an external force application module configured to apply an external force to the three-dimensional cellular neural network model;

[0093] a cell current calculation module configured to calculate a cell current of a cell in the three-dimensional cellular neural network model according to the applied external force;

[0094] a state voltage calculation module configured to calculate a state voltage of the cell according to the cell current of the cell;

[0095] a displacement calculation module configured to calculate a displacement of the cell according to the state voltage of the cell;

[0096] a deformation simulation module configured to perform deformation simulation on the three-dimensional cellular neural network model according to the displacement of the cell.

[0097] Specifically, the state voltage calculation module is specifically configured to calculate the state voltage of the cell according to the following formula:

[0098]

[0099] wherein, v xijk (t) is the state voltage of the cell C(i,j,k) at time t, R x is the cell resistance, I ijk (t) is the cell current of the cell C(i,j,k) at time t, C(a,b,c) is an adjacent cell of the cell C(i,j,k), N r (i,j,k) is a set of adjacent cells of the cell C(i,j,k), A (i，j，k；a，b，c) (t) is the feedback coefficient of the cell C(i,j,k) and the adjacent cell C(a,b,c) at time t, v yabc (t) is the output voltage of the adjacent cell C(a,b,c) at time t.

[0100] Specifically, the displacement obtaining module is specifically configured to obtain the displacement of the cell according to the following formula:

[0101] ||Δl ijk (t)||=β||v xijk (t)||||E ijk ||

[0102] wherein, Δl ijk (t) is the displacement of the cell C(i, j, k) at coordinate (i, j, k) in the three-dimensional cell neural network model at time t, ||Δl ijk (t)|| is the numerical value of Δl ijk (t); ||v xijk (t)|| is the numerical value of v xijk (t); the direction of Δl ijk (t) is the same as the direction of v xijk (t); β is a cell displacement coefficient; E ijk is the elastic modulus of the cell C(i, j, k), and ||E ijk || is the numerical value of E ijk .

[0103] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage media, etc.) containing computer-usable program code.

[0104] The present application is described with reference to flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in a flow or multiple flows and / or blocks Figure 1 The functions specified in a flow or multiple flows and / or blocks

[0105] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0106] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that are executed on the computer or other programmable apparatus provide steps for implementing the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0107] The above embodiments of the present application have been described in conjunction with the accompanying drawings, but the present application is not limited to the above described embodiments, and the above described embodiments are merely illustrative, not restrictive, and those skilled in the art can make many modifications without departing from the spirit and scope of the present application, and these modifications are also within the scope of the present application.

[0108] The basic principles and main features of the present application and the advantages of the present application have been shown and described above. It should be understood by those skilled in the art that the present application is not limited to the above described embodiments, and the above described embodiments and descriptions in the specification are merely illustrative of the principles of the present application, and various changes and modifications can be made without departing from the spirit and scope of the present application, and these changes and modifications are also within the scope of the present application. The scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A soft tissue deformation simulation method based on cellular neural networks, characterized in that, include, Construct a three-dimensional cellular neural network model; An external force is applied to a three-dimensional cellular neural network model, and the cell currents of the cells in the three-dimensional cellular neural network model are obtained based on the applied external force. Determine the state voltage of a cell based on its cellular current; The cell displacement is determined based on the cell's state voltage; Deformation simulation of a three-dimensional cellular neural network model is performed based on cell displacement. The method of determining the state voltage of a cell based on its cellular current includes, in, for Coordinates in a time-space 3D cellular neural network model Cell State voltage, For cell resistance, for Time Cell Cellular currents, For cells neighboring cells, For cells The collection of neighboring cells, for Time Cell and neighboring cells Feedback coefficients of interaction, for Neighboring cells at all times ; output voltage; The The calculation formula is as follows: = ； in, The energy transfer coefficient, and Cells and neighboring cells The elastic modulus, express The value, for Neighboring cells at all times Output voltage direction and Time Cell Starting point, neighboring cells The cosine of the angle between the directions of the vectors ending at the endpoint.

2. The soft tissue deformation simulation method based on cellular neural networks according to claim 1, characterized in that, The process of determining the cell currents in the three-dimensional cell neural network model based on the applied external force includes, ； definition for Coordinates in a time-space 3D cellular neural network model Cell Cellular currents, for The value; cell When it is a point of force application, define Time Cell The direction of cell current and Time Cell The directions of the external forces acting on them are consistent; for Time Cell stress The value; cell When it is a non-force-bearing point, It is 0.

3. The soft tissue deformation simulation method based on cellular neural networks according to claim 1, characterized in that, It also includes, the The calculation formula is as follows: ； in, for The value; The initial interval length between adjacent cells in a three-dimensional cellular neural network model. The value; The cross-sectional area of ​​the cell The value; This is the cell displacement coefficient; For cells elastic modulus The value.

4. The soft tissue deformation simulation method based on cellular neural networks according to claim 1, characterized in that, The The calculation formula is as follows: ； in, for Neighboring cells at all times State voltage, direction and The directions are the same. This is the saturation voltage of the cell.

5. The soft tissue deformation simulation method based on cellular neural networks according to claim 1, characterized in that, The method of determining cell displacement based on cell state voltage includes, Among them, the definition for Coordinates in a time-space 3D cellular neural network model Cell displacement, for The value; for Time Cell State voltage, for The value; direction and The directions are the same; This is the cell displacement coefficient; For cells The elastic modulus, for The value.

6. A soft tissue deformation simulation device based on cellular neural networks, characterized in that, include: Modules for building three-dimensional cellular neural network models; External force application module, used to apply external force to the three-dimensional cellular neural network model; The cell current calculation module is used to calculate the cell current of cells in a three-dimensional cell neural network model based on the applied external force. The state voltage determination module is used to determine the state voltage of a cell based on its cellular current. The displacement calculation module is used to calculate the cell displacement based on the cell's state voltage. The deformation simulation module is used to perform deformation simulation on a three-dimensional cellular neural network model based on cell displacement. The state voltage calculation module is specifically used to calculate the cell's state voltage using the following formula: ； in, for Coordinates in a time-space 3D cellular neural network model Cell State voltage, For cell resistance, for Time Cell Cellular currents, For cells neighboring cells, For cells The collection of neighboring cells, for Time Cell and neighboring cells Feedback coefficients of interaction, for Neighboring cells at all times ; output voltage; The The calculation formula is as follows: = ； in, The energy transfer coefficient, and Cells and neighboring cells The elastic modulus, express The value, for Neighboring cells at all times Output voltage direction and Time Cell Starting point, neighboring cells The cosine of the angle between the directions of the vectors ending at the endpoint.

7. The soft tissue deformation simulation device based on cellular neural networks according to claim 6, characterized in that, The displacement calculation module is specifically used to calculate the cell displacement using the following formula: Among them, the definition for Coordinates in a time-space 3D cellular neural network model Cell displacement, for The value; for Time Cell State voltage, for The value; direction and The directions are the same; This is the cell displacement coefficient; For cells The elastic modulus, for The value.

Images