EP4193217A1 - Verbesserte berechnung ophthalmischer linsen - Google Patents
Verbesserte berechnung ophthalmischer linsenInfo
- Publication number
- EP4193217A1 EP4193217A1 EP21755448.4A EP21755448A EP4193217A1 EP 4193217 A1 EP4193217 A1 EP 4193217A1 EP 21755448 A EP21755448 A EP 21755448A EP 4193217 A1 EP4193217 A1 EP 4193217A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- model
- parameters
- ophthalmic lens
- lenses
- surface model
- 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
Classifications
-
- G—PHYSICS
- G02—OPTICS
- G02C—SPECTACLES; SUNGLASSES OR GOGGLES INSOFAR AS THEY HAVE THE SAME FEATURES AS SPECTACLES; CONTACT LENSES
- G02C7/00—Optical parts
- G02C7/02—Lenses; Lens systems ; Methods of designing lenses
- G02C7/024—Methods of designing ophthalmic lenses
- G02C7/028—Special mathematical design techniques
Definitions
- the present invention relates to a method for defining a surface model, a method for determining at least one surface of at least one ophthalmic lens using a surface model, and a corresponding manufacturing method. Furthermore, the invention relates to corresponding computer program products and devices.
- One task when calculating ophthalmic lenses is to calculate the shape of the surfaces of an ophthalmic lens or a pair of lenses and their location (ie orientation and position) in relation to one another in such a way that they have certain geometric (e.g. at certain points thicknesses specified by the lens) and optical properties which are adapted to the later wearing situation (e.g. adaptation to the eye or pair of eyes looking through the ophthalmic lenses in optical and possibly also physiological terms).
- ophthalmic lenses examples include contact lenses and spectacle lenses such as single vision lenses and contact lenses, multifocal spectacle lenses and multifocal contact lenses and spectacle lenses with a variable refractive power (e.g. progressive lenses).
- contact lenses and spectacle lenses such as single vision lenses and contact lenses, multifocal spectacle lenses and multifocal contact lenses and spectacle lenses with a variable refractive power (e.g. progressive lenses).
- the type of adjustment is usually limited to the ametropia of the eyes, the refractive index of the material used for the lenses, and the size and/or shape of the frame of the spectacle lens.
- the ametropia can include sphere, cylinder and axis, possibly an addition or a near refraction, and/or a prismatic prescription.
- individualized lenses individual parameters such as the orientation and distance of the lenses to the viewing eye (given via Centering parameters), distance or position of the eye rotation points to the glasses and/or to each other, individual object distances in certain visual points (e.g. reference points) in the lens, and the individual position of these visual points in the glass are added as further parameters.
- further parameters can be added, such as the visual situation specifically intended for the lens, the visual behavior (e.g. the interaction of head and eye deflection), biometric parameters that describe the eye (e.g. world front errors including higher orders (e.g. as a Zernike coefficient set ), pupil sizes and/or positions at different gaze directions, eye length, curvatures and positions of the refracting surfaces of the eye, refractive index of the media) or other parameters specific to the intended wearer.
- the visual situation specifically intended for the lens e.g. the interaction of head and eye deflection
- biometric parameters that describe the eye e.g. world front errors including higher orders (e.g. as a Zernike coefficient set )
- pupil sizes and/or positions at different gaze directions e.g. as a Zernike coefficient set
- eye length e.g. as a Zernike coefficient set
- the shape of the surfaces of an ophthalmic lens is often described as a free-form surface, which can be parameterized, for example, by a set of so-called arrow heights.
- Other local representations such as a spline representation, or non-local representations, such as a Zernike decomposition, are also possible. If one of the surfaces is a comparatively simple surface, e.g. a sphere, only its curvature or a so-called base curve can be specified.
- surface parameters e.g. by superimposing a specific surface defined for a product with another surface in order to calculate the prescription in one set reference point or measuring point.
- an optimization is often used, for which the shape and position of the starting surfaces and one or more target functions to be optimized are necessary.
- Such starting areas can either be kept constant for a large number of ordering parameters, or several starting areas belonging to different ordering parameter sets can be interpolated and/or extrapolated.
- a disadvantage of the conventional calculation methods for ophthalmic lenses is that with an increasing number of ordering parameters, either a direct calculation (e.g. by superimposition) is no longer possible or is associated with a loss of quality, or that the calculation time for an optimization run becomes longer and longer, since the Optimization evaluated objective functions become more complex.
- Calculations of ophthalmic lenses are usually carried out anew each time, even if ophthalmic lenses have identical or very similar order parameters and the surfaces of the ophthalmic lenses calculated in this way are therefore identical or similar.
- Such calculations are typically performed in the manufacture of ophthalmic lenses (to define the surfaces to be manufactured, in the design of ophthalmic lenses, or to check the manufacturability of an ophthalmic lens based on geometric properties of the surfaces). They are also used in consultations (e.g. at the optician's) to explain the optical and geometric properties of an individualized or personalized ophthalmic lens to the future wearer of such a lens (e.g. position of the zones of clear vision and thicknesses for progressive lenses). In order to keep waiting times short, the calculations must be completed in a short time in this special application.
- One object of the invention is to reduce the computing effort required for the calculation of ophthalmic lenses while at the same time limiting the memory requirement. This enables a faster calculation of ophthalmic lenses that uses less computing capacity and is therefore more economical.
- This object is achieved by a computer-implemented method, a corresponding device and a corresponding computer program product for defining a surface model, a computer-implemented method for determining at least one surface of at least one ophthalmic lens, a corresponding device and a corresponding computer program product, a method for and a corresponding device for manufacturing an ophthalmic lens having the features specified in the respective independent claims.
- a computer-implemented method for determining a surface model for calculating at least one surface of at least one ophthalmic lens e.g. a contact lens or a spectacle lens
- a set of ordering parameters for the at least one ophthalmic lens and/or from sizes dependent on the ordering parameters (such as sizes derived from the ordering parameters).
- the "calculation of at least one area of an ophthalmic lens” in the sense of the present application includes the calculation of at least a part of an area or an area piece.
- “calculating at least one area of an ophthalmic lens” means calculating at least part of the area or calculating the entire area.
- the at least one ophthalmic lens can be a single lens. It is also possible to calculate one or both lenses of a pair of ophthalmic lenses. For example, at least one pair of ophthalmic lenses (lens pair) can be calculated with the surface model, which includes lenses intended for the right and left eye of a person.
- the set of order parameters can contain order values for both lenses of the pair of lenses (e.g. for the left and the right spectacle lens of a pair of spectacle lenses as well as binocular order data).
- the at least one surface calculated according to the surface model can be described parametrically by at least one parameter.
- calculating the area using the area model includes calculating at least one parameter of the area (area parameter) from the order parameters or from sizes (auxiliary sizes), which are dependent on the order parameters (such as sizes, which are derived from the order parameters).
- the surface can be described, for example, by the curvature or the main curvatures in at least one point, e.g. a reference point of the ophthalmic lens, as well as by a surface normal and possibly the orientation of the main cuts. Furthermore, it is possible to describe the surface using a local representation, such as a spline representation or a polynomial representation with the appropriate coefficients, or a non-local representation, such as a Zemike generation with the appropriate coefficients.
- a local representation such as a spline representation or a polynomial representation with the appropriate coefficients
- a non-local representation such as a Zemike generation with the appropriate coefficients.
- the calculation of the area using the area model includes the calculation of the versine of the area in a plurality of grid points from the order parameters or from sizes (auxiliary sizes), which are dependent on the order parameters (e.g. are derived from the order parameters).
- one of the surfaces of the ophthalmic lens and/or the arrangement of this surface in relation to the other surface of the ophthalmic lens is calculated using the surface model from the order parameter set.
- the other surface may be a predetermined surface, e.g., a spherical surface with a predetermined curvature, which may be dependent on ordering parameters (such as a known base curve system).
- ordering parameters such as a known base curve system.
- both surfaces of the ophthalmic lens and/or their mutual arrangements are calculated using the surface model from the order parameter set.
- the at least one surface from the set of ordering parameters for the ophthalmic lens is preferably calculated using the surface model directly (ie without iteration) or with a few iteration steps, such as less than 30, 25, 15, 10, 5 or 3 iteration steps.
- the ophthalmic lens, the at least one surface of which is calculated using the surface model from a specific order parameter set, is referred to in the context of the application as a lens calculated according to the surface model from this order parameter set.
- the lens calculated according to the surface model can be one of the lenses of a pair of lenses. In this case, one or both lenses of the pair of lenses can be calculated according to the surface model.
- the ophthalmic lens can, for example, be a spectacle lens, for example an individualized and/or personalized spectacle lens.
- the orientation of the spectacle lens in front of the user's eye is taken into account when calculating or optimizing the spectacle lens, for example.
- the orientation of the spectacle lens can be characterized, for example, by the forward tilt, the frame lens angle, the pupillary distance, the corneal vertex distance and/or other parameters.
- an adjustment of the perceived design is made in order to meet a person-related purpose of using a pair of glasses.
- the spectacle lens can be, for example, a single-vision spectacle lens, a multifocal spectacle lens or a progressive spectacle lens.
- the surface model can be a parameterized model.
- the surface model can include at least one variable parameter.
- the surface model can have at least one constant parameter (e.g. the position(s) of the evaluation points in the glass).
- the procedure for determining a surface model includes the steps:
- a training data set comprising a multiplicity of order parameter sets, each of which contains values of at least some of the parameters required for ordering at least one ophthalmic lens;
- Obtaining the surface model for calculating at least one surface of at least one ophthalmic lens comprising:
- Determining optimized values for the model parameters of the at least one surface model can include:
- Optimizing the values of the model parameters of the at least one surface model (optimizing the parameterization of the surface model) with the aim of minimizing or maximizing a target function for the model parameters of the at least one surface model that depends at least on the model parameters and on the target values provided.
- the target function for the model parameters for each of the order parameter sets contains at least one term which assumes a minimum or maximum if the provided target value of the at least one property of the at least one ophthalmic lens for the respective order parameter set corresponds to the value of the same property at least one with the surface model given Values of the model parameters of the surface model for the corresponding order parameter set of calculable or calculated lenses match.
- the plurality of order parameter sets can each contain values of at least some of the parameters required for ordering the pair of lenses.
- the at least one property may include a binocular property of the pair of lenses.
- At least one lens pair can be calculated with the surface model, which includes lenses intended for the right and left eye of a person.
- the order parameter sets contain order parameters for one at a time right and a left lens.
- the properties for which target values have been provided may include at least one binocular property dependent on the left and right lens area data.
- the target values for the at least one binocular property can include values whose calculation includes at least one property (eg a surface property) at a first location of a first lens of a pair of lenses and the same property at a second location of a second lens of the same pair of lenses.
- an initial complexity of the surface model can also be provided.
- an initial number of model parameters can be specified.
- the complexity of the surface model can also be optimized or adjusted.
- Optimizing the complexity of the surface model can include, for example, varying the number of model parameters and/or regularization.
- the provision of at least one surface model parameterized by model parameters can include the provision of at least two surface models of different complexity, the complexity of a surface model comprising one or more of the following variables:
- the procedure may also include:
- a validation data set comprising a multiplicity of order parameter sets, each of which contains values of at least some of the parameters required for ordering at least one ophthalmic lens; and Providing at least one target value of at least one property of the at least one ophthalmic lens for each of the order parameter sets in the validation data set.
- Obtaining a surface model for calculating at least one surface of at least one ophthalmic lens may further include:
- providing in the sense of the present application includes “determining”, “transmitting”, “receiving”, “reading out”, “removing from a memory, a database and/or a table”, “receiving”, etc.
- the order parameter sets required to calculate at least two other different ophthalmic lenses or pairs of lenses are provided. It is advantageous to use more than 10, 100, 1000, 10000, 100000 or 1000000 order parameter sets required for calculating the further ophthalmic lenses when specifying the surface model.
- the order parameter sets preferably cover a large, preferably the entire, range in which ophthalmic lenses can later be ordered (see, for example, the limits for refraction specified by manufacturers of ophthalmic glasses, individual frame parameters when ordering spectacle lenses, other parameters of the glasses such as freely selectable object distances, and other order parameters).
- the order parameter sets can range from refraction values, e.g. -20 dpt to +20 dpt for sphere and -8 dpt to +8 dpt for cylinder.
- An ordering parameter set can include one, several or all ordering parameters required for ordering an individual ophthalmic lens or a pair of ophthalmic lenses. Examples of order parameters can also be found in the current standards for spectacle lenses (cf. e.g. EU Directive 93/42/EEC on medical products).
- An order parameter set can include at least one of the following order parameters:
- ophthalmic lens such as material (possibly with refractive index of the lens), desired thickness of the lens, coating, etc.;
- - refraction values such as sphere and/or cylinder with axis and/or addition and/or near refraction and/or prism with base;
- ophthalmic lens a purpose of use of the ophthalmic lens, eg reading, computer work, sports, etc.; - physiological parameters or characteristics of the future wearer of the ophthalmic lens;
- Biometric parameters or properties of the eye or eyes of the future wearer such as the position of the center of rotation of the eye, individual structure of the eye, pupil diameter, individual measurement of a wave front, etc.;
- the parameters for individualizing an ophthalmic lens can, for example, characterize the orientation of the ophthalmic lens in front of the eye of a lens wearer.
- the parameters for personalizing an ophthalmic lens can characterize an adaptation of the perceived design, for example to suit a personal use of glasses. These parameters for individualizing and/or personalizing an ophthalmic lens can be mapped, for example, in a design characteristic, the location of the reference points, length of progression, etc.
- variables dependent on the order parameters can be, for example, the refractive index of the material, mechanical properties of the material, mechanical properties of the coating, thickness distribution of the ophthalmic lens, desired distribution of the residual astigmatism in the position of use, desired distribution of the refractive error in position of use, etc.
- the ordering parameter sets required to define the surface model can, but do not necessarily have to, relate to ophthalmic lenses that have already been ordered, calculated, or manufactured. Depending on the embodiment of the surface model, it can be advantageous that the ophthalmic lenses have already been ordered, calculated or manufactured. It is therefore also possible that the order parameter sets are only within the permissible limits of the order parameters. So the refraction can be in the delivery range for ophthalmic lenses, but the glasses themselves can never be ordered, calculated or to have been manufactured.
- redundant orders in an order data record can be removed before the determination, if necessary, in order to reduce the number of data records (e.g. in the case of effects that are ordered very frequently).
- the order data record can be stratified for the same reason, but still ensure a high coverage of the order parameter area with order data records.
- At least one target value of at least one predefined property of the at least one ophthalmic lens is provided for defining the surface model for each of the order parameter sets in the training data set.
- the target values for the different order parameter sets can be different or at least identical for some of the order parameter sets.
- target value includes a desired or required value of at least one property of an ophthalmic lens, e.g., a spectacle lens or a contact lens.
- the target value can include multiple values or be a combination of multiple values.
- the target value can, for example, be taken from a database and/or calculated using predefined optimization algorithms.
- the at least one predefined property of the ophthalmic lens can be, for example, an optical or geometric property of an ophthalmic lens or of a pair of lenses comprising the ophthalmic lens.
- the at least one specified property can be a physical property of the lens, such as arrow height(s), curvature(s) or variables derived therefrom, such as surface astigmatism, surface refractive index, etc..
- the at least one specified property can also be a "Indirect" property, i.e. a property that is associated with at least one model (e.g. object distance model, eye model, usage position model, etc.) arises. Examples of indirect properties are residual astigmatism,
- the property can be one of the following properties:
- optical variables or properties of at least one surface or the ophthalmic lens such as refractive index or refraction errors (preferably in the position of use), astigmatism or residual astigmatism (preferably in the position of use) vertical and / or horizontal prism (preferably in the position of use), aberrations of a higher order (preferably in the position of use), etc.
- the optical quantities or properties can be specified, for example, in power vector form;
- Distribution of optical variables or properties and/or their gradients of at least one surface or the ophthalmic lens such as distribution of the refractive error, the vector components and/or the amount and/or the axis of the astigmatism or residual astigmatism in the position of use, the prism , the base of the prism, the vector components of the prism, or distribution of quantities derived therefrom.
- Distribution means both the optical property and its gradient understood as a function of the spatial position (eg (xy) position) on the ophthalmic lens, as well as the frequency distribution of these variables in the sense of a probability distribution;
- the - width of the zones of good vision e.g. the sun, in which the residual astigmatism and/or the refractive error are less than 1 dpt, preferably less than 0.75 dpt or 0.5 dpt;
- geometric parameters or properties of the ophthalmic lens such as central thickness of the lens, edge thickness of the lens, thickness of the coating, diameter of the lens, mass of the lens, etc.;
- the binocular property can be a property in whose calculation at least one property (e.g. a surface property, a geometric property, an optical property, a property of visual perception, etc.) at a first point of a first lens of a pair of lenses (such as a pair of spectacle lenses ) and the same property occurs at a second point of a second lens of the same pair of lenses.
- the at least one binocular property of the pair of lenses can be, for example, the deviation or the difference in at least one optical quantity or property between the first and the second lens of a pair of lenses.
- Exemplary binocular properties of a pair of lenses are the deviation or the difference of the horizontal and/or vertical prism (minus the prescribed prism difference) in corresponding visual points (e.g. in at least one reference point, e.g. in the prism reference point) between the right and the left lens of a pair of lenses, the difference in magnification between the left and right lenses, the base curve difference between the left and right lenses, the location the design points or their difference, properties of the visual perception of the future wearer of the ophthalmic lenses calculated with the surface model, etc.
- the parameters or properties of the pair of lenses can be the deviation of at least one geometric parameter between the first and the second lens of a pair of lenses, such as center thickness variation, edge thickness, coating thickness, front surface curvature, etc.;
- a wearer of the ophthalmic lenses calculated with the surface model has a feeling of discomfort related to the quality of vision and/or the posture;
- the target values can be values (such as target values) of the at least one ophthalmic property of lenses that have already been calculated or are to be calculated or manufactured using a known method (e.g. by minimizing or maximizing a known target function in an iterative optimization method) for the different sets of order parameters. It is also possible to obtain the target values using measured values from ophthalmic lenses that have already been manufactured.
- the ordering parameters of the lenses that have already been manufactured or are to be manufactured are preferably already known, at least in part.
- a target value of the at least one property of the at least one ophthalmic lens As a target value of the at least one property of the at least one ophthalmic lens, a measured value of the at least one property of an ophthalmic lens that has already been manufactured or a value that is determined or can be determined from one or more measured values of ophthalmic lenses that have already been manufactured can be set. It is also possible to set a desired value of an ophthalmic lens to be manufactured as the target value. For example, a target value for the mean surface power of a surface of the ophthalmic lens as a function of position on the lens, given a known index of refraction, can be determined from sagittal measurements of the glass surface (via the mean curvature of the surface).
- a target value for the astigmatism of a spectacle lens in the wearing position can be determined from the sagittal measurements, the forward inclination, the frame lens angle, lens length and lens height (order values) and the position of the ocular pivot point (order values or model assumptions).
- one or more surfaces of these lenses belonging to the order parameter sets can be used in addition to the order parameter sets for a large number of ophthalmic lenses (base lenses).
- base lenses ophthalmic lenses
- the ophthalmic lenses used to define the surface model are also referred to as base lenses within the scope of the present application.
- the base lenses can be lenses that have been calculated or optimized and optionally manufactured using a known method.
- the target functions dependent on these order parameter sets and required for the calculation of ophthalmic lenses according to the prior art or for the calculation of the base lenses and/or their derivations according to the surfaces can also be used.
- the latter can be understood, for example, as a change in the target function when the arrow heights change.
- the target values can thus correspond to the target values, which are included in these target functions.
- target functions and/or their derivatives according to the areas are preferably available in such a way that they can be evaluated for any areas.
- the target functions and/or their derivatives are typically evaluated as a function of a suitable parameterization of the surfaces of the ophthalmic lenses.
- the surfaces of a large number of base lenses are used to define the surface model, it can be advantageous if these surfaces have already been calculated. It can also be an advantage if the base lenses have also been manufactured, since their calculation already had a purpose and was used for Definition of the surface model is reused without additional computing capacity would be consumed.
- the method described here can also easily be carried out with measured surfaces (and distances between the surfaces) of ophthalmic lenses instead of with calculated surfaces.
- other properties such as the refractive index can also be measured.
- properties and target values in relation to the training data set.
- properties and target values can also be determined or calculated for other order parameter sets (e.g. validation and test data set) in the same or similar way.
- the same properties with corresponding target values do not necessarily have to be used for the training, validation and test data sets. However, it can often be easier to use the target values of the same properties in relation to different datasets.
- the surface model can be any model, such as a machine learning based model.
- Machine learning algorithms are described e.g. in Jeremy Watt, Reza Borhani, Aggelos Katsaggelos: Machine Learning Refined: Foundations, Algorithms, and Applications, Cambridge University Press, 2020.
- the surface model can be described by suitably defined model parameters, the model parameters being used together with at least some or at best all order parameters and/or variables derived therefrom for calculating the surface or surfaces of the ophthalmic lens.
- the surface model is such is constructed so that the surfaces generated by the surface model are a continuous or even continuously derivable function of the ordering parameters in order to ensure continuity of the design of the ophthalmic lenses with regard to these ordering parameters.
- a surface model defined by model parameters can be designed as a regression model or contain a regression model.
- the regression model receives at least a part or preferably old order parameters and/or variables derived therefrom as input variables and uses them to calculate one or more surfaces of an ophthalmic lens or of a pair of ophthalmic lenses.
- the coefficients of the regression model represent at least part of the model parameters of the surface model.
- the surface model or parts thereof can also be designed as a classification model in addition or as an alternative to being designed as a regression model, or can contain a classification model. If, for example, only certain lens diameters are available for the blanks during the production of spectacle lenses, and the diameter of the lens blank is to be selected, this can be done using a classification model.
- a classification model can, for example, calculate the probability of the suitability of the available glass blank diameters for the ophthalmic lens to be manufactured, so that ultimately the glass blank can be selected that has the highest probability and is therefore best suited for the ophthalmic lens to be manufactured.
- a classification model can also be used to calculate the optimal base curve and/or the optimal diameter of the ophthalmic lenses (e.g. spectacle lenses) during the production of ophthalmic lenses (e.g. spectacle lenses) by calculating a probability for the suitability of the available base curves , and the base curve with the highest probability for the production or calculation of the ophthalmic lens (eg the spectacle lens) is selected.
- the optimal base curve and/or the optimal diameter of the ophthalmic lenses e.g. spectacle lenses
- the optimal diameter of the ophthalmic lenses e.g. spectacle lenses
- the surface model has a sufficiently high number of model parameters, eg more than 10, 30, 50, 100, 500, or 1000, 10000, 100000 or even more model parameters.
- the determination of the surface model can consist of setting the model parameters of the surface model based on the order parameter sets (e.g. order parameter sets of the base lenses) in such a way that any ophthalmic lenses, their surface or surfaces can be calculated using the surface model from their order parameter set, differ only slightly from ophthalmic lenses based on predetermined criteria, which can be calculated from the same order parameter set using predetermined methods according to the prior art or have already been calculated.
- the order parameter sets e.g. order parameter sets of the base lenses
- the regression model used in the surface model or as a surface model can be a linear regression model, which typically simplifies the calculations required to define the surface model, since the model parameters can be determined from a linear system of equations.
- non-linear regression model instead of a linear regression model.
- Such a model is more flexible and can depict more complex relationships between order parameters and the surface(s) of the ophthalmic lenses.
- nonlinear optimization algorithms are typically used for this purpose, which do not necessarily converge to the global optimum of the model parameters.
- Neural networks which also include deep neural networks, can be used as non-linear regression models, but other non-linear regression models known from the field of machine learning can also be used. These regression models, eg the neural network, can be trained using the order parameter sets provided in the training data set with the associated target values.
- the surface model can also be a combination of a linear and/or non-linear regression model, a classification model and/or a neural network.
- a reduction in the complexity of the surface model and savings in computing time and/or consumption of resources are conceivable as a result of the combination. Exemplary combinations are:
- classification model before regression model and/or neural network
- the model with the smallest complexity that still generates sufficiently good surfaces is selected from the set of regression models with different complexity.
- regression model before classification model before regression model and/or neural network Based on the order data, the geometry of the lens is approximately determined with the help of a regression model of low complexity. The subsequent classification model determines the glass blank from the order data and the approximation geometry. The subsequent regression model determines the final lens from the order data and the glass blank.
- An advantage of this procedure is a reduction in the complexity of the regression model for determining the lens.
- auxiliary variables auxiliary variables
- At least one surface model parameterized by model parameters is provided or specified, with which—with given values of the model parameters—at least one surface of at least one ophthalmic lens can be calculated from at least one order parameter set and/or from variables dependent on an order parameter set.
- an initial parameterization and an initial complexity of the surface model can be provided or specified.
- the provision of an initial parameterization of the surface model can include the provision of initial values for the model parameters of the surface model.
- the provision of an initial complexity of the surface model can include the definition or specification of an initial number of model parameters of the surface model.
- the final model parameters and, if necessary, complexity are determined or defined using a suitable optimization method. The final model parameters thus form an optimal set of model parameters.
- the model parameters of the surface model are preferably defined in such a way that they represent an optimal set of model parameters which minimizes or maximizes a specified target function for the model parameters.
- An optimal set of model parameters can be found using common mathematical optimization algorithms (eg a simple gradient descent, conjugated gradient descent, stochastic gradient descent or similar algorithms). If a neural network is used as the regression model, a backpropagation algorithm can be used to minimize the target function, which in itself is only a gradient-based algorithm adapted for this type of model.
- mathematical optimization algorithms eg a simple gradient descent, conjugated gradient descent, stochastic gradient descent or similar algorithms.
- the target function for the model parameters for each of the ordering parameter sets can contain at least one term which assumes a minimum or maximum if the provided target value of the at least one property of the at least one ophthalmic lens for the respective ordering parameter set with the value of the same property at least one with the surface model given values of the model parameters of the surface model for the corresponding order parameter set of calculable or calculated lenses.
- the objective function can contain a single term, a sum of multiple terms, or a weighted sum of multiple terms.
- different objective functions can be used for the model parameters.
- different target functions can be used, depending on (i) whether there are already existing areas (e.g. areas calculated according to a conventional method) for the order parameter sets and these areas are to be used in the target function, or (ii) whether the areas corresponding to the order parameter sets would have to be calculated first, or (iii) whether the objective function should not be defined using areas.
- the gradient of the target function with regard to the parameters in the parameterization of the surface output by the surface model can be calculated quickly, e.g. as an analytical function.
- the optimization of the parameterization and, if necessary, the complexity of the surface model can be done in such a way that the deviations of the values of at least one property of ophthalmic lenses, the at least one surface of which is calculated using the surface model from the sets of order parameters, and the corresponding (possibly order parameter-dependent) target values this property are minimized.
- the deviation of the at least one value of the at least one specified property of a lens calculated or calculable according to the surface model from a specific order parameter set from the at least one target value of this property for the same order parameter set can be quantified in different ways.
- the difference or a convex or concave function of the difference (e.g. a square, an amount, or their negative) between the at least one value of the at least one specified property calculated or calculable according to the surface model from a specific order parameter set can be used as a measure for this deviation Lens and the at least one target value for this property can be used for the same order parameter set.
- the at least one term of the target function for the model parameters can be the difference or a convex or concave function of the difference between the at least one value of the at least one specified property of the lens, the at least one surface of which is calculated or can be calculated according to the surface model for an ordering parameter set, and which include at least one target value for this property for the same order parameter set.
- a convex function can be used in particular when minimizing the objective function. When maximizing the objective function, a concave function can be used.
- Such a function is, for example, the ratio of the at least one value of the at least one specified property of the lens calculated or calculable according to the surface model from a specific order parameter set and the at least one target value for this property for the same order parameter set.
- Other functions such as a logarithmic function of the ratio, are also possible.
- target functions for the at least one optical property which are used in a conventional optimization or calculation of at least ophthalmic lenses.
- One or more terms of the target function for the model parameters can form or be interpreted as a target function for optimizing or calculating at least one ophthalmic lens for given order parameter sets, with the target function being evaluated for the different order parameter sets.
- the target function can be calculated in a known manner from the actual value of the at least one optical property (evaluated for an ophthalmic lens whose at least one surface has been calculated or can be calculated according to the surface model from a specific part parameter set and/or variables derived therefrom) and the corresponding target value depend. It is also possible to use different objective functions for different order parameter sets.
- such a target function can be the sum running over a large number of viewing points or the mean value of the squares of the deviations of the refractive error and the astigmatism calculated over a large number of viewing points from the respective target values calculated directly from the order parameter set, which in turn sums over the order parameter sets or is averaged.
- the target function for the model parameters can contain several terms which, based on different properties (e.g. optical and/or geometric properties, direct and/or indirect properties), quantify or calculate the differences between the surfaces of the ophthalmic lenses calculated with the surface model and the corresponding target values. describe.
- the additional properties eg the additional optical and/or geometric properties, direct and/or indirect properties
- Exemplary properties are vector components or the amount of residual astigmatism in the usage position, refractive error in the usage position, deviation of the minimum or maximum lens thickness from the corresponding order value of the lens thickness, etc..
- the objective function can also contain terms that quantify or describe the difference between the binocular properties of two pairs of ophthalmic lenses (a pair that was calculated using the surface model and a pair that was calculated using methods according to the prior art). .
- the objective function can contain at least one term which contains the design differences between ophthalmic lenses with different ordering parameters. Such an additional term therefore no longer refers only to a single ophthalmic lens, but to the differences between two or more ophthalmic lenses that are adjacent in the ordering area, and reflects advantageous properties of a product containing a large number of ophthalmic lenses.
- the similarity (but not necessarily the sameness) of the perceived design of progressive lenses across different refractions can be formulated as a target function, so that the desired design only has to be specified for a single effect, which e.g. occurs particularly frequently, and itself the designs resulting from other effects without having to be specifically stated.
- the advantage of such terms, which include the differences between two or more ophthalmic lenses, is that it is often difficult to specify a surface design that is constant over the ordering parameter range, since it may not be constant due to other more fundamental principles (e.g., Minkwitz's theorem). .
- the objective function may include a weighted or unweighted sum of the terms established for each of the order parameter sets over all order parameter sets in the training data set. Instead of a sum is possible the mean or to form the median. It is also possible to use more complex functions, such as non-linear functions, instead of a sum.
- optimization algorithms e.g., a simple gradient descent, conjugated gradient descent, stochastic gradient descent, or similar algorithms
- optimizing the values for the model parameters may include regularizing the objective function used in optimizing the model parameters.
- an optimization method is used to optimize the parameters of the surface model, which requires gradients of the target function (e.g. in the backpropagation algorithm for neural networks as a surface model, or e.g. gradient descent), these can be calculated analytically, numerically or with the help of combined analytical and numerical methods .
- the gradients of the objective function are backpropagated through the network in the backpropagation step instead of the commonly used residuals (the commonly used residuals are the gradient of a commonly used quadratic objective function).
- target functions for determining the model parameters.
- other target functions such as a validation target function or a target function for testing the surface model (test target function)
- test target function can also be set up or specified in the same way.
- the objective function can be normalized. For example, the sum of properties of the ophthalmic lens and ophthalmic lens (pairs) can be divided by the number of ophthalmic lenses (pairs) (each based on the number of lenses (pairs) in the training, validation or test data set, depending on whether the objective function is used to determine the model parameters, to validate the model or for testing. Setting the complexity of the surface model
- the complexity of a surface model can include one or more of the following:
- Adjusting or optimizing the complexity of the surface model can also provide:
- a validation data set comprising a multiplicity of order parameter sets, which each contain values of at least some of the parameters required for ordering at least one ophthalmic lens
- a target value of at least one property of the at least one ophthalmic lens for each of the order parameter sets in the validation data set include.
- obtaining the surface model for calculating the at least one surface of at least one ophthalmic lens can include:
- the validation objective function depends on the provided objective values.
- the validation target function contains at least one term for each of the order parameter sets in the validation data set, which assumes a minimum or maximum if the provided target value of the at least one property of the at least one ophthalmic lens for the respective order parameter set has the value of the same property at least one optimized with the surface model at given Values of the model parameters of the surface model for the corresponding order parameter set calculable or calculated lens matches.
- the validation target function can be constructed in the same way or similarly to the target function for the model parameters. However, it is possible to use different objective functions.
- the target function(s) for the optimization of the model parameters and the validation target function(s) can be the same terms dependent on the respective target values (ie related to the training data set or the validation data set). contain.
- the terms of the target function for the optimization of the model parameters, which contain the regularization parameter or parameters can be omitted for the calculation of the validation target function.
- the same can also be achieved by setting the regularization parameter(s) in such a way that the corresponding terms do not contribute to the validation objective function (e.g. by setting the regularization parameters to 0). In all of this, of course, the corresponding target values must be replaced with those based on the validation data set and not on the training data set.
- the squares of the differences in the arrow heights can be minimized in the optimization (i.e. the corresponding properties would be the arrow heights at given evaluation points on the ophthalmic lens) and in the validation the squares of the difference in the power of the ophthalmic lens and the corresponding target values (here the property would be the effect, e.g. as a power vector, sphere/cylinder/axis, or one or more components of the effect).
- the model parameters can be selected in such a way that the surfaces output by the surface model are compatible with the surfaces already calculated (target surfaces ) match as closely as possible. Possible criteria for this can be defined in the target function for the model parameters.
- the target function contains a term that is a sum of a convex function (e.g. the square) of the differences in the versine heights of the surfaces calculated or calculable by the surface model with a given model parameter set and the target surfaces.
- the sum runs point by point over all pairs of versine heights of target surfaces and the surfaces calculated or calculable with the surface model, as well as over the base lenses or the order parameter sets.
- the areas of the base lenses that have already been calculated together with the associated ordering parameters and/or variables derived from them can be viewed as a training data set.
- a sum weighted differently over the sagitta of an ophthalmic lens can also be used.
- the weights at points on the ophthalmic lens that are to be assessed as particularly critical can be higher than at other points.
- the weights can be higher in the area of the ophthalmic lens that is looked through more frequently in order to ensure high optical quality (e.g. in the case of spectacle lenses, the region of the tubular round spectacle lens that is inside the frame after grinding, or in varifocal lenses (e.g. the range in which the residual astigmatism is below a certain threshold, eg 0.5 dpt).
- the target function for the model parameters of the surface model can also contain terms that penalize a deviation of the diameter calculated by the surface model or implicitly resulting from the calculation results (e.g. because a curvature is too high) (e.g. the target function increases very strongly, if the calculated diameter is smaller than the nominal diameter).
- the already calculated areas of the base lenses and the areas calculated by the surface model are given in different parameterizations (e.g. the arrow heights are specified for different point grids), it is advantageous to convert the already calculated areas of the base lenses to the parameterization output by the surface model, e.g Interpolation.
- the parameterization of the surfaces can also be adjusted in the opposite direction, or a completely different parameterization can be selected (e.g.
- the target function for the model parameters can additionally or alternatively contain other terms which, based on optical and/or geometric properties, quantify the differences between the surfaces of the ophthalmic lenses calculated with the surface model and the base lenses.
- the optical or geometric properties can also depend on the order parameters (e.g. vector components or the amount of residual astigmatism in the position of use, refractive error in the position of use, or deviation of the minimum or maximum glass thickness from the corresponding order value of the glass thickness).
- these terms consist of sums or weighted sums of the pointwise differences of the optical quantities of the lenses calculated with the surface model and the base lenses, which in turn are summed over all base lenses.
- the points over which summation is carried out can be specified by a grid of evaluation points of the ophthalmic lenses, or by a grid of viewing directions.
- a (weighted) sum other functions, e.g. non-linear functions, can be used.
- the objective function can also contain terms that quantify the difference between the binocular properties of two pairs of ophthalmic lenses (a pair calculated using the surface model and a pair calculated using a conventional optimization method).
- objective functions are used for the model parameters that do not depend on the areas of base lenses. This can be the case, for example, if not enough of these surfaces have been calculated to determine the model parameters, or if there are still no correspondingly calculated surfaces for the order parameter sets of the base lenses.
- the difference between specified properties of the surfaces calculated with the surface model and desired target values of these properties can be calculated.
- Possible properties are the optical properties mentioned above, e.g. optical properties (e.g. distribution of the refractive error, the vector components and/or the magnitude and/or the axis of astigmatism or residual astigmatism in the position of use, the prism, the prism base, the vector components of the prism, or distribution of quantities derived from it), geometric properties, binocular properties or properties of the visual perception of the future wearer of the ophthalmic lenses calculated with the surface model.
- optical properties e.g. distribution of the refractive error, the vector components and/or the magnitude and/or the axis of astigmatism or residual astigmatism in the position of use, the prism, the prism base, the vector components of the prism, or distribution of quantities derived from it
- geometric properties e.g. distribution of the refractive error, the vector components and/or the magnitude and/or the axis of astigm
- the target function for the model parameters can contain terms that represent a weighted deviation of the properties of the ophthalmic lens calculated or calculable with the surface model from its desired course.
- the weighting can be used, for example, to control compliance with the effect required by the standards (e.g. high weighting in the reference points for spectacle lenses).
- the target function for the model parameters can also contain terms that have a minimum at the desired central thickness or thickness distribution at the edge. Desired mechanical properties such as actual or simulated breaking strength can also be mapped in a term of the objective function.
- the objective function can be one of the objective functions described above.
- all examples of target functions that do not depend on the areas of the base lenses can be used for the model parameters from the previous section by replacing the corresponding property of the base lenses with suitably selected target values that may depend on the ordering parameters.
- the target function for the model parameters can also contain terms that are already used as target functions in the optimization of ophthalmic lenses with the aid of customary optimization methods.
- the target function of the model parameters then contains terms which sum the target functions of optimization methods over the further ophthalmic lenses or form their mean value.
- Defining the model parameters by optimizing a target function for the model parameters that does not depend on the areas of base lenses therefore corresponds to the simultaneous optimization of a number of ophthalmic lenses that are obtained from the order parameter sets by calculation and/or optimization be able.
- the model parameters of the surface model that control the course of the surfaces are varied in order to calculate the sum of the individual target functions for the optimization of each to determine the Model parameters used to minimize ophthalmic lenses that can be calculated or optimized from the order parameter sets.
- the objective function for the model parameters of the surface model can also contain terms containing design differences between ophthalmic lenses with different order parameters. The advantages of such terms have already been discussed above.
- the areas calculated by the area model are a continuous or even continuously derivable function of the order parameters and/or quantities derived from them, it can be expected that if the complexity of the area model is suitably set, there will be only slight differences between the areas calculated by the area model and the areas which were calculated by optimizing the same objective functions using conventional optimization methods.
- the so-called regularization can also be used, in which further terms weighted with one or more different factors are added to the target function of the parameters of the surface model. Typically these terms are quadratic terms in the model parameters.
- other powers can also be used (e.g. the absolute value of the model parameters can be used), or other functions of the model parameters can also be used instead of the model parameters themselves (e.g. differences in the spline coefficients of neighboring splines of a representation of the surface of the ophthalmic lens ).
- model parameters of the surface model In order to check the quality of the surface model after the model parameters have been set, it is advisable not to compare the model parameters of the surface model with all the available data sets (i.e. at least the data sets containing the order parameters, any sizes derived from them and any associated calculated surfaces). train, but to use part of the datasets for validation or for final testing of the model.
- the validation of the set model complexity and the final test of the surface model can be done using the same objective function for model parameters that was used to set the model parameters of the surface model, but preferably without the terms originating from the regularization.
- target functions that differ from the target function for determining the model parameters.
- the procedure for specifying a surface model can therefore include the following steps:
- the method for defining a surface model can include the following steps:
- the purpose of validation can be to determine a suitable model architecture (here also called model complexity) or a suitable value of the regularization parameter(s).
- the purpose of testing can be to check the trained and selected model to avoid overfitting.
- the function evaluated during validation and testing can be the same, eg the validation function (also called test objective function) described above.
- the validation objective function usually does not contain any additional terms that contain regularization parameters.
- the validation target function can be evaluated with fixed model parameters of the surface model on the test data set and compared with the values of the validation target function evaluated on the validation data set and/or the target function for the model parameters (but without regularization terms).
- the test is successful if the value of the validation objective function evaluated on the validation dataset and the value of the validation objective function evaluated on the test dataset are of similar size. How much they actually differ from each other depends on the data (including the number, which has a strong influence if no objective functions normalized to the number of data are used) in the respective validation or test data set and on the underlying model.
- Values of validation and test target functions that are not normalized to the number of data can only be compared with one another if the Test and validation data sets contain the same amount of data. If these functions are divided by the number of data, one obtains normalized target functions that can also be compared if the test and validation data set contain different numbers of data points.
- the values of the target functions to be compared i.e. value of the validation target function when evaluated with the test data set and the validation data set
- do not differ too much from each other e.g. the absolute value of the difference between two values of the target function should be less than be a predetermined threshold.
- the value for such a threshold depends strongly on the type of objective function used and should be a small fraction (e.g. 0.3 to 0.01 times) of the variation of the validation objective function when evaluating with different models or different values of the regularization parameter(s). s) (e.g. maximum value - minimum value).
- these can be divided beforehand by the number of ophthalmic lenses (pairs) in the respective data sets and be normalized in this way.
- sizes can be used as input sizes of the surface model which depend on the ordering parameters, such as size, which have been derived from the ordering parameters.
- the surface model contains a regression model, for example, or if the surface model consists of a regression model, it can be advantageous that the input variables of the regression model, from which the surfaces are calculated, contain one or more variables calculated from the order parameters (auxiliary variables) in addition to or instead of the order parameters ) include.
- auxiliary variables are:
- the desired thickness distribution in one or more points (e.g. at the edge) of the ophthalmic lens which also depends, for example, on the material and/or layer or coating of the ophthalmic lens or the optical and/or mechanical properties of the material and/or the layer or .Coating may be dependent;
- optical and/or mechanical properties of the material of the ophthalmic lens e.g. refractive index, modulus of elasticity, thermal expansion coefficient
- Optical and/or mechanical properties of the coating of the ophthalmic lens e.g. thickness distribution, modulus of elasticity, thermal expansion coefficient.
- those parameters are preferably to be selected which have a great influence on the surfaces of the ophthalmic lenses or which are expected to have a great influence on the surfaces.
- the input layer of the neural network is assigned the ordering parameters and/or auxiliary variables calculated from them.
- the weights of the neural network represent at least part of the model parameters.
- the output layer can represent the entire calculated area or parts of the calculated area of the ophthalmic lens (e.g. as arrow heights in a defined grid or grid with the desired resolution that may still have to be set).
- the neural network can also contain one or more hidden layers.
- model parameters e.g. their number
- auxiliary variables calculated from the order parameters to the input layer in addition to or instead of the order parameters (cf. the section "Calculating variables derived from order parameters" for exemplary auxiliary variables which are used as input variables for a surface model be able). It can also be advantageous to design the neural network in such a way that one or more of these auxiliary variables are at least approximately represented in the network or are set during the training of the network. The model parameters that are not used as weights of the neural network can also be included in the calculation of the auxiliary variables.
- the output layer may also be advantageous to limit the size of the neural network by having the output layer represent only a relatively coarse grid of the ophthalmic lens area (e.g. a grid of only 10 x 10 or 20 x 20 arrow heights).
- a relatively coarse grid of the ophthalmic lens area e.g. a grid of only 10 x 10 or 20 x 20 arrow heights.
- it makes sense to interpolate the arrow heights output by the neural network on a grid with a higher resolution (e.g. using linear or bicubic interpolation on a grid of e.g. 100 x 100 Points) and, if necessary, post-optimize with a few steps of an optimization method according to the prior art.
- this post-optimization requires only a few iterations for convergence—provided the neural network was trained with the results of an optimization method with the same target function.
- the determined or defined surface model with the optimized model parameters and optionally the optimized complexity can be suitably stored and then made available for the calculation of ophthalmic lenses from an ordering parameter set.
- the surface model or a part of the surface model (such as the model parameters) can be stored, for example, in a suitable memory, such as in a database.
- At least part of the order parameter sets required to define the surface model and/or the associated target values can also be stored in the memory.
- the target values can be, for example, surface values or quantities derived therefrom of ophthalmic lenses (base lenses) that have already been calculated at least in part using common methods.
- the target values can also be target values that are used in a target function for optimizing an ophthalmic lens (e.g. a Objective function according to the prior art).
- the surface model determined as described above can be further modified. Accordingly, the method can include modifying the surface model. Exemplary modifications are appending additional layers to a neural network or embedding the surface model in another function that interpolates or transforms, for example, versine heights of the surfaces (e.g. transformation of a neural network into a support vector machine, a decision tree or any other regression model).
- less than 90%, preferably less than 50%, 20%, 10%, 5%, 2% or 1%) of the computing power or computing time used to calculate the areas is used for iterative changes in the areas of the ophthalmic lenses. This can be advantageous compared to conventional iterative algorithms, since the calculation effort remains the same with each iteration, but the changes in the areas decrease with each iteration.
- the surface model does not necessarily require the provision of an initial surface (start surface) as input, which is included in the calculation as an input variable and is modified during the optimization process.;
- Hybrid calculations of the areas with a subsequent correction are possible. Such hybrid calculations can ultimately have a shorter optimization time due to a better starting surface;
- the method according to the above aspect also has advantages in the development of series of ophthalmic lenses (both in relation to individual ophthalmic lenses and pairs of ophthalmic lenses), since the surface model can be used unchanged if the distinguishing between two different series Ordering parameters (e.g. refractive index) or quantities dependent on them (e.g. base curve systems) were present in the training data set and varied.
- series Ordering parameters e.g. refractive index
- quantities dependent on them e.g. base curve systems
- Another application of the method according to the above aspect is the interpolation between different series of ophthalmic lenses (eg between different products such as series of progressive lenses intended for different uses, or eg between progressive lenses and single vision lenses).
- the order data record only has to be expanded by a size that corresponds to a series of ophthalmic lenses use.
- a series of ophthalmic lenses i is represented by the tuple ( ⁇ 1 , i , ⁇ 2,i ,..., ⁇ N,i ), where ⁇ j,i is the Kronecker delta symbol.
- Interpolation between different series of ophthalmic lenses can then be made possible by choosing the values s j between 0 and 1, where the sum of the s j is 1.
- Such a tuple is then used together with a common order data set to calculate the at least one surface of ophthalmic lenses using a correspondingly trained surface model.
- a second aspect of the invention relates to a computer-implemented method and a corresponding device for determining at least one of the surfaces of one or more ophthalmic lenses using a previously defined surface model from the order parameters and/or from variables derived therefrom.
- the term "determining" within the meaning of the present application includes determining or calculating at least one surface of one or more ophthalmic lenses.
- the procedure includes:
- the surface model can be the surface model described above, ie a surface model that has been determined or ascertained using the method described above.
- the surface data of the at least one surface are preferably determined directly (ie not iteratively) or with a few iteration steps from the order parameter set provided, such as with fewer than 30, 25, 15, 10, 5 or 3 iteration steps.
- this leads to a significant reduction in the computation time required to create a surface or lens for any one Calculate order parameter set.
- the surface model can be used directly to calculate the at least one lens surface.
- a function of the surface model can be used, such as a function that approximately carries out the calculation with a surface model determined according to the invention.
- a function can be generated as part of a simplification of the surface model, e.g. by combining neurons of a neural network that have similar activation patterns, or as part of another transformation of the surface model used.
- the method also includes determining other variables relevant to the production of the surfaces (e.g. existing diameters and type of blanks from which the glasses are to be produced), so that the surfaces calculated in this way either require no further optimization or only with a comparatively small amount computational effort must be corrected.
- the method for determining at least one surface of at least one ophthalmic lens can also carry out a correction of the at least one surface calculated with the surface model, the correction being an optimization of the surface calculated with the surface model and/or an overlay with an overlay surface and/or a correction of production-related deviations of the surfaces or the optical properties of the ophthalmic lens and/or an extension of the area to the diameter required for manufacturing the ophthalmic lens.
- a correction of the at least one surface calculated with the surface model the correction being an optimization of the surface calculated with the surface model and/or an overlay with an overlay surface and/or a correction of production-related deviations of the surfaces or the optical properties of the ophthalmic lens and/or an extension of the area to the diameter required for manufacturing the ophthalmic lens.
- the method for determining at least one surface of at least one ophthalmic lens can include storing surface data of the at least one surface calculated with the surface model and optionally corrected and/or expanded.
- the area data can optionally be stored together with at least part of the order parameter set used to determine the area data.
- the area data can be stored, for example, on a suitable data carrier or in a storage device.
- the storage device can also be a computer or data cloud.
- the method for determining at least one surface of at least one ophthalmic lens can also include the transmission of surface data of the at least one surface calculated with the surface model and optionally corrected and/or expanded to an external unit, such as a manufacturer of ophthalmic lenses, a manufacturing unit, a Manufacturing device, etc.
- the surface data can optionally be transmitted together with at least part of the order parameter set used to determine the surface data.
- the method for determining at least one surface of at least one ophthalmic lens can include checking the at least one surface calculated with the surface model for fulfillment of desired or required properties and storing the information about the fulfillment or non-fulfilment of the required properties together with at least part of the Determination of the surface data using the order parameter set used and/or with the at least one surface calculated with the surface model and, if necessary, corrected and/or expanded and/or with at least one value of the desired or include required properties.
- the method for determining at least one surface of at least one ophthalmic lens can include adapting the model parameters of the surface model after determining and/or storing each or after a predetermined number of surfaces calculated and optionally corrected with the surface model.
- the at least one surface of the ophthalmic lens calculated using the surface model can be corrected further if, for example, the ophthalmic lens calculated using the surface model does not meet at least one desired or required optical and/or geometric property.
- the ophthalmic lens calculated with the surface model can be checked for compliance with the desired or required properties, e.g. by exceeding or falling below suitably selected threshold values.
- a correction of the ophthalmic lenses calculated with the surface model based on these surfaces can consequently take place independently of whether the fulfillment of the required properties of the ophthalmic lens is checked or not.
- the correction can be made, for example, using state-of-the-art methods. For example it is possible to calculate the surfaces of the ophthalmic lens in a post-optimization, which includes a few optimization steps of a common optimization method for ophthalmic lenses.
- the surfaces output by the surface model can be used as the starting point (so-called starting surface) for post-optimization.
- a calculation can also be carried out using a second surface model.
- the second surface model can include a regression model, for example.
- the surface or surfaces of the ophthalmic lens can be checked again after the correction to ensure that the desired or required properties are met. If the ophthalmic lens does not meet the desired or required properties, the manufacture of the lens can be stopped in order to manually review the order. This ensures that the production or even delivery of unsuitable lenses is prevented. In particular, failed post-optimizations can be discovered in this way.
- an ophthalmic lens has finally been post-optimized after its calculation by the surface model, its surface or surfaces can be stored as a new data record in a database together with the ordering parameters and possibly other variables derived therefrom.
- This data set can be used to improve the surface model by determining or specifying the surface model again taking into account the newly added data.
- the database for the method according to the invention is thus constantly growing.
- the renewed definition of the surface model can also be done with only a part of the data records stored in the database.
- the quality of the surfaces calculated with the surface model can be improved with each post-optimized lens.
- a correction of the surface or surfaces calculated with the surface model can be omitted if, for example, the deviation of the order parameter set of the ophthalmic lens to be calculated and/or variables derived from it from the order parameter set required to define the surface model, which is most similar to the order parameter set of the lens to be calculated , is less than a predefined threshold.
- the deviation can be measured using a suitably defined distance measure, which can be selected in such a way that the sensitivity of the surface or surfaces is taken into account by the ordering parameters.
- the deviation can be described or quantified using the squares of the arrow height differences or the differences in the desired properties of two ophthalmic lenses with different ordering parameters.
- the model parameters of the surface model can be checked and/or modify continuously or at regular intervals. If ophthalmic lenses need to be re-optimized or re-optimized, the data required for the surfaces of the ophthalmic lenses are generated and can be used together with the associated ordering parameters for adjusting the model parameters. Optimization algorithms that only use part of the data, such as stochastic gradient descent or limited memory BFGS, can preferably be used for adaptation. However, optimization algorithms that require the complete data set can also be used.
- the model parameters of the surface model can be checked and/or adjusted, for example, after each or after a predetermined number of newly calculated or post-optimized ophthalmic lenses. In the simplest case, this number can be constant.
- Another possibility is to only make an adjustment when a fixed proportion (e.g. 10%) of the data already used to determine the model parameters has been re-optimized or post-optimized. It may also make sense to adapt the model parameters of the surface model if computing time is available (e.g. if few ophthalmic lenses have to be calculated).
- a fixed proportion e.g. 10%
- computing time e.g. if few ophthalmic lenses have to be calculated.
- model parameters of the surface model are to be adjusted by using only new data sets from the post-optimization, then it is also possible to use the learning rate (i.e. the strength of the adjustment of the model parameters in an adjustment step) when adjusting the model parameters proportionally to the proportion of the number of new datasets based on the total number of datasets used for training. In this way, the learning rate is reduced with each adjustment and ensures the convergence of the model parameters.
- the learning rate i.e. the strength of the adjustment of the model parameters in an adjustment step
- a hybrid method for determining the at least one of the surfaces of one or more ophthalmic lenses can be provided. include determining the at least one of the surfaces using a previously defined surface model from the order parameters and/or from variables derived therefrom and a subsequent correction or recalculation of the at least one surface determined using the surface model.
- the correction can be any of the corrections described above.
- Z(F_m(a_m)) be the quality of the surfaces F_m(a_m) calculated with the surface model, measured using the target function Z, where a_m is the computational effort of the surface model.
- Z(F_n(a_n;F_m(a_m))) correspond to the quality of the areas F_n(a_n;F_m(a_m)) calculated starting from the areas F_m(a_m) by recalculation with the computational effort a_n.
- the derivatives on both sides have a negative sign, since the target function becomes smaller with increasing computational effort. It is possible to restrict the range of ordering parameters in which the surface model is used to the range of frequent orders in order to keep the overall complexity of the model small and thus minimize the computational effort averaged over all ordered ophthalmic lenses. This can be useful, for example, if certain sets of order parameters occur rarely and others very often.
- the surface model can only be used for lenses with standard individual parameters and standard designs in the main order area (e.g. sphere between -4 dpt and +4 dpt; amount of the cylinder under 2 dpt; addition between 1.5 dpt and 2.5 dpt). In the remaining area of the order parameters, the methods according to the state of the art can then be used.
- the complexity of the surface model can, for example, also be restricted in such a way that only ordering parameters with the greatest influence on the surface or surfaces of the ophthalmic lens are processed.
- the influence of the other ordering parameters on the areas can be corrected afterwards with a recalculation or reoptimization.
- the ordering parameters shown in the surface model or used to define the surface model are reduced to prescription values and centering parameters including forward tilt, frame lens angle and corneal-vertex distance for an ophthalmic lens.
- the calculation of the surface or surfaces of the ophthalmic lens remains more efficient than the previously practiced optimization based on a standard starting surface.
- the area in which the surface model is used for the calculation can be defined using common classification algorithms, such as logistic regression or support vector machines.
- the area of the order parameters, in which the area model is used to calculate the areas can also be continuously expanded: If, for example, outside the area of the order parameters, in which the area model is used for the calculation, sufficient areas finally calculated using state-of-the-art methods for are available, the model parameters of the surface model can be trained again. The extended area can then also be recalculated as suggested above.
- the division, in which areas of the ordering parameters which surface model is used, can be determined experimentally.
- the goal of an optimal distribution can be the fast and resource-saving calculation of the be ophthalmic lenses.
- a computer program product which, when loaded into the memory of a computer and executed on a computer, causes the computer to carry out a method according to one of the above aspects.
- the computer can also be a computer system.
- the method according to one of the above aspects can be carried out using a correspondingly designed device.
- a fourth aspect of the invention relates to a device for determining a surface model for calculating at least one surface of at least one ophthalmic lens (e.g. a contact lens or a spectacle lens) at least from a set of ordering parameters for the ophthalmic lens and/or from sizes dependent on the ordering parameters.
- the device comprises a computing device which is designed to carry out the method according to the first aspect of the invention.
- the device for determining a surface model comprises in particular: a device for providing a training data set comprising a large number of order parameter sets, which each contain values of at least some of the parameters required for ordering at least one ophthalmic lens; a device for providing at least one target value of at least one predetermined property of the at least one ophthalmic lens for each of the order parameter sets in the training data set; a device for providing at least one surface model parameterized by model parameters, with which - given values of the model parameters - at least one surface of at least one ophthalmic lens can be calculated from at least one order parameter set and/or from variables dependent on an order parameter set; and a computing device which is designed to obtain or determine the surface model for calculating at least one surface of at least one ophthalmic lens, wherein obtaining or determining the surface model comprises:
- Determining optimized values for the model parameters of the at least one surface model includes, for example:
- a fifth aspect of the invention relates to a device for determining at least one of the surfaces of one or more ophthalmic lenses using a previously determined surface model from the ordering parameters and/or from variables dependent on the ordering parameters.
- the device is designed to carry out the method for determining at least one of the surfaces of at least one ophthalmic lens according to one of the above aspects.
- the device for determining at least one surface of an ophthalmic lens comprises in particular: a device for providing an order parameter set for the at least one ophthalmic lens; • a device for providing a surface model for calculating at least one surface of at least one ophthalmic lens from a set of ordering parameters for the at least one ophthalmic lens and/or from the Sizes dependent on ordering parameters; and a computing device which is designed to determine surface data for the at least one surface of the at least one ophthalmic lens using the surface model from the order parameter set provided.
- the surface model can be a surface model that was determined or ascertained using the method according to one of the aspects described above.
- a sixth aspect of the invention relates to a data set comprising surface data of at least one surface of at least one ophthalmic lens, wherein the at least one surface has been determined using the method for determining at least one of the surfaces of at least one ophthalmic lens according to one of the above aspects.
- the data set may be stored or may have been stored permanently or non-permanently on a suitable data carrier or in a storage device, such as a database, a computer or data cloud, etc.
- the method described above for determining at least one of the surfaces of one or more ophthalmic lenses using a surface model and the corresponding device can be used in the manufacture of ophthalmic lenses, e.g. to define the surfaces to be manufactured, in the design of ophthalmic lenses, or to Check the manufacturability of an ophthalmic lens based on the geometric properties of the surfaces.
- the method for determining at least one of the surfaces of one or more ophthalmic lenses can be part of a manufacturing or manufacturing method for ophthalmic lenses.
- the device for producing an ophthalmic lens comprises in particular: a device for determining at least one surface of at least one ophthalmic lens according to one of the aspects described above; a manufacturing device for manufacturing the ophthalmic lens having the at least one surface.
- the above-mentioned devices for providing, determining, specifying or calculating data (such as quantities derived from order parameters, model parameters, target values, area data, weightings, etc.) and/or for evaluating functions, such as target functions, can be replaced by suitably configured or Programmed data processing devices (in particular specialized hardware modules, computers or computer systems, such as computers or data clouds) can be implemented with appropriate computing units, electronic interfaces, memories and data transmission units.
- the devices may further comprise at least one graphical user interface (GUI), preferably interactive, allowing a user to view and/or enter and/or modify data.
- GUI graphical user interface
- the above-mentioned devices can also have suitable interfaces that enable data (such as order parameter sets, model parameters, target values, surface data, etc.) to be transmitted or input or read out.
- the devices can also include at least one memory unit, for example in the form of a database, which stores the data used, such as order parameter sets, target values, area data, weightings, etc.
- the production device can, for example, include at least one CNC-controlled machine for the direct processing of a blank according to the determined optimization specifications.
- the ophthalmic lens can be manufactured using a casting process.
- the finished ophthalmic lens preferably has a simple spherical or rotationally symmetrical aspherical surface and a surface determined using the surface model according to one aspect of the method according to the invention.
- the simple spherical or rotationally symmetrically aspheric surface is the front surface (i.e., the object-side surface) of the ophthalmic lens.
- the surface calculated with the surface model as the front surface of the ophthalmic lens. Both surfaces of the ophthalmic lens and/or their mutual arrangement can also be determined with the aid of the surface model.
- the invention offers a use of an ophthalmic lens produced according to the production method according to the invention in a predetermined average or ideal usage position of the spectacle lens in front of the eyes of a specific wearer to correct ametropia of the wearer.
- FIG. 2 shows an exemplary method for calculating an ophthalmic lens with the aid of a parameterized surface model
- FIG. 3 shows a further exemplary method for calculating an ophthalmic lens with the aid of a surface model and with a correction
- FIG. 4 shows a further exemplary method for calculating an ophthalmic lens with the aid of a surface model and with optional correction
- FIG. 5 shows an exemplary method for defining a surface model with the aid of ophthalmic lenses that have already been calculated
- FIG. 6 shows an exemplary method for defining a surface model without ophthalmic lenses that have already been calculated
- FIG. 7 shows an exemplary division of a data set comprising a large number of order parameter data sets in a training data set, a validation data set and a test data set;
- FIG. 8A shows the correlation of the center thickness of spectacle lenses, which were calculated using a first exemplary surface model, and the center thickness of test spectacle lenses;
- Fig. 8B is a histogram of the frequency of center thickness residuals
- 8C shows the correlation of the back surface curvature of the first principal section of spectacle lenses, which were calculated using the first surface model, and the back surface curvature of the first principal section of test spectacle lenses;
- 8E shows the correlation of the back surface curvature of the second principal section of spectacle lenses, which were calculated using the first surface model, and the back surface curvature of the second principal section of test spectacle lenses;
- 8F is a histogram of the frequency of the residuals of the back surface curvature of the second major section
- 8G shows the correlation of the vertex power in the first principal meridian of spectacle lenses, which were calculated using the first surface model, and the vertex power in the first principal meridian of test spectacle lenses;
- 8H shows a histogram of the frequency of the deviations of the vertex power calculated according to the first surface model in the first principal section from the vertex power in the first principal section of test spectacle lenses;
- 8J shows a histogram of the frequency of the deviations of the vertex power calculated according to the first surface model in the second principal section from the vertex power in the second principal section of test spectacle lenses;
- FIG. 9A shows the correlation of the center thickness of spectacle lenses, which were calculated using a second exemplary surface model, and the center thickness of test spectacle lenses;
- Fig. 9B is a histogram of the frequency of center thickness residuals
- 9C shows the correlation of the back surface curvature of the first principal section of spectacle lenses, which were calculated using the second surface model, and the back surface curvature of the first principal section of test spectacle lenses;
- 9E shows the correlation of the back surface curvature of the second principal section of spectacle lenses, which were calculated with the aid of the second surface model, and the back surface curvature of the second principal section of test spectacle lenses;
- 9G shows the correlation of the vertex power in the first principal section of spectacle lenses, which were calculated using the second surface model, and the vertex power in the first principal meridian of test spectacle lenses;
- 9H shows a histogram of the frequency of the deviations of the vertex power calculated according to the second surface model in the first principal section from the vertex power in the first principal section of test spectacle lenses;
- 9I shows the correlation of the vertex power in the second principal section of spectacle lenses, which were calculated using the second surface model, and the vertex power in the second principal section of test spectacle lenses;
- 9J shows a histogram of the frequency of the deviations of the vertex power calculated according to the second surface model in the second principal section from the vertex power in the second principal section of test spectacle lenses;
- a conventional method for calculating an ophthalmic lens Li for an order parameter set d i usually includes the following steps:
- the optimization is usually carried out iteratively by minimizing or maximizing a target function, in which target values for at least one property of the lens (e.g. an optical property are included.
- the target function is usually evaluated for a specific parameterization of the area to be calculated. The parameters of the Areas are changed until specified criteria are met.
- the procedure includes the steps: S1-1: providing ordering data comprising an ordering parameter set d k for the ophthalmic lens or the ophthalmic lenses;
- S1-2 Calculation/optimization of at least one surface of the lens(es) using a surface model
- the at least one surface of the lens can be calculated using the surface model directly, not iteratively, or according to an iterative method with a few iteration steps. This significantly reduces the time required to calculate the area.
- the surface model can be a model that has been defined according to one of the aspects and embodiment variants described above.
- the surface model can be defined parametrically, for example, with the model parameters (parameters of the parametric representation of the surface model) being used together with at least some of the ordering parameters and/or quantities derived therefrom to calculate the surface or surfaces of the ophthalmic lens.
- the surface model can be a linear or non-linear regression model.
- the non-linear regression model can be a neural network.
- the procedure includes the steps:
- Provision of order data comprising an order parameter set d k for the ophthalmic lens or for the ophthalmic lenses;
- the surface model can be defined using existing order parameter sets with the associated target values. For this purpose, an initial complexity and an initial parameterization can be determined or specified.
- the model parameters can subsequently be determined using an optimization method in which the model parameters are changed iteratively. The aim of the optimization method is for the surfaces and/or their properties output by the surface model for different sets of ordering parameters to correspond as well as possible to the target values for the same sets of ordering parameters.
- the optimization of the parameterization and possibly the complexity of the surface model can be done, as described above, by minimizing or maximizing a target function for the model parameters, the target function preferably being evaluated over all order parameter sets in the training data set.
- the target function contains at least one term which is dependent on the deviation of the value or values determined for each order parameter set in the training data set of at least one specified property of an ophthalmic lens calculated according to the surface model from the at least one target value of this property for the same order parameter set.
- the objective function for the model parameters may include the following term: whereby:
- Z i (j) denotes the j th value of the at least one property Z of a lens calculated according to the surface model for the ith order parameter set;
- Z i (j) Soll designates the j th target value of the at least one property Z for the ith order parameter set;
- g z (j) denotes the weight of the jth value of the at least one property Z.
- the jth value of the at least one property Z of the lens can be determined using the current parameterization or the current model parameters of the surface model.
- the jth value of the at least one property Z of the lens can be, for example, the value of this property at the jth assessment point of the lens.
- the function f i can be a target function, for example, which is used for the optimization of ophthalmic lenses according to a conventional method and which is evaluated for the current parameterization or the current parameter of the surface model.
- One or more target values Z i (j) target can be equal to 0.
- the astigmatism in the position of use can have the target value of 0 dpt at one or more evaluation points of the lens.
- the above function f i and/or its derivatives according to the surfaces can then be evaluated using all order parameter sets in the training data set, with the evaluation being carried out as a function of the parameterization of the surface model.
- a weighted or non-weighted sum f can be formed and evaluated from the functions f i determined for all order parameter sets and/or their derivations according to the areas.
- N is the number of order parameter sets (e.g. the number of order parameter sets in a training set); and denotes the weight of the i-th term for the i-th order parameter set, which is identically equal to 1 for an unweighted sum.
- the model parameters are modified and the target function f is re-evaluated. This is repeated iteratively until the predetermined criteria are met.
- the surface model with the model parameters determined in this way can be suitably stored and, as described above, used to calculate new ophthalmic lenses.
- FIG. 3 shows another exemplary method for calculating an ophthalmic lens or a pair of ophthalmic lenses to order data using a surface model and direct calculation.
- the method is similar to the method shown in FIG. 1 and also includes a correction of the surface calculated with the surface model.
- the correction of the area can be one of the corrections described above.
- the procedure includes the steps:
- S3-1 providing order data comprising an order parameter set d k for the ophthalmic lens or for the ophthalmic lenses;
- S3-4 obtaining the surfaces of the ophthalmic lens(es) L k to be manufactured for the order parameter set d k .
- the correction of the surface or surfaces calculated with the surface model can be one of the corrections described above. Due to the optimal starting area for the post-calculation or post-optimization, such a correction usually requires one or only a few iterations. The total computing time can be considerable as a result be reduced.
- FIG. 4 shows another exemplary method for calculating an ophthalmic lens or a pair of ophthalmic lenses to order data using a surface model and direct calculation and with optional correction.
- the procedure includes the steps:
- S4-1 providing order data comprising a multiplicity of order parameter sets d k for the ophthalmic lens or for the ophthalmic lenses;
- S4-2 Calculation/optimization of the ophthalmic lens(es) using a surface model (directly, preferably not iteratively or with a few iterative steps);
- S4-4 carrying out a correction of the surface(s) calculated with the surface model (post-calculation/post-optimization with one or only a few iterations) if a correction is necessary;
- FIG. 5 shows an exemplary method for defining a surface model using ophthalmic lenses that have already been calculated.
- the procedure includes the steps:
- Provision of an order data set ⁇ d i ,L i ⁇ comprising a large number of order parameter sets ⁇ d i ⁇ and a large number of already calculated surfaces of a large number of ophthalmic lenses L i , and division of the data set into training (possibly validation -) and test data set.
- the ophthalmic lenses L i provided are lenses which have been calculated for the order parameter sets in the order data set using a known calculation or optimization method.
- the lenses provided can be lenses which have each been optimized using a target function f i ;
- S5-3 Optimization of the parameterization and, if necessary, the complexity of the surface model (iterative) with the aim that the surface model reproduces the training (and possibly validation data set) as well as possible.
- the optimization is based on a target function G for model parameters.
- ophthalmic lenses BL i or their surfaces are calculated using the surface model for the order data d t .
- the surfaces or lenses calculated with the surface model are compared using the target function G with the surfaces of the provided lenses L i or with the provided lenses.
- the sum of the target functions G(BL i ,L i ) over i is minimized;
- the above method can also be carried out with measured surfaces and/or distances between the surfaces of lenses that have already been manufactured instead of with calculated surfaces.
- FIG. 6 shows an exemplary method for defining a surface model without ophthalmic lenses already calculated.
- the procedure includes the steps:
- S6-3 Optimization of the parameterization and, if necessary, the complexity of the surface model with the aim that the lenses L i calculated from the surface model minimize the sum of the target functions f i running over the training (and possibly validation data set) for the optimization of individual glasses .
- the target functions f i can be target functions known from the prior art;
- S6-5 Obtaining optimized parameters of the surface model and, if necessary, an optimized complexity of the surface model to provide for a direct calculation of ophthalmic lenses from order data.
- each of the ordering parameter sets d k may include one or more ordering parameters needed to order a single ophthalmic lens or a pair of ophthalmic lenses.
- ordering parameters can be found in the current standards for spectacle lenses (cf. eg EU Directive 93/42/EEC on medical products).
- ordering parameters and the values derived from them as well as with regard to further details, reference is made to the above explanations in the relevant sections. All of the features, design variants and/or advantages described there apply analogously to the above examples.
- a first example relates to a lens calculation using a regression model.
- the back surface curvature and the center thickness of single vision lenses are calculated directly from the order values sphere and cylinder of the refraction using a surface model designed as a regression model.
- the parameters and the complexity of the regression model are determined based on data from already calculated ophthalmic lenses.
- the lens diameter is specified as 65mm.
- the starting point for the calculation is a data set with a total of 825 lenses already calculated using state-of-the-art methods, the sphere and cylinder of which vary in steps of 0.25 dpt.
- the dataset was divided into a training dataset of 425 lenses, a validation dataset of 192 lenses, and a test dataset of 208 lenses, as shown in FIG.
- the subdivision is according to a certain pattern (see Fig. 7), and not random as usual, since the data form an equidistant grid into sphere and cylinder. However, it is possible to randomly divide the initial data set into a training data set, a validation data set, and a test data set.
- the calculation of the curvature of the front surface KVFL is carried out using tables in which the base curves are tabulated depending on the sphere, cylinder.
- the curvature of the front surface could also be calculated using a classification model from sphere, cylinder, and possibly from the refractive index of the material and possibly the information about the glass blanks available for production. However, this has been intentionally omitted in this example for the sake of clarity.
- a regression model is then determined, which in this example is formed with the aid of spline functions.
- Cubic splines are used in each of K1, K2 and KVFL, and their linear interaction term is also used.
- the node points of the splines are distributed equidistantly in the value range of K1, K2 and KVFL, ie the nodes form a grid with equal distances in the respective parameters K1, K2 and KVFL.
- the spline coefficients represent the model parameters.
- the sum of the squares of the residuals (i.e. the differences in the variable to be calculated minus the corresponding value in the data set) was used as a measure of how well the model describes the available data.
- the square deviation of the center thickness in mm and the two curvatures of the back surface in dpt were minimized in separate adjustments.
- the curvatures of the back surface are in this example in dioptres based on the Index of refraction 1.525 given.
- the model parameters contained in Table 2 were determined by minimizing the deviation of the actual center thickness from the center thickness calculated using splines (model parameters in mm) using the training data set.
- the spline basis functions are numbered below with a multi-index, where 0 means that the spline function in the order parameter or in the derived variable is constantly equal to 1, and higher indices in increasing order correspond to the cubic spline basis functions that are in the value range of the respective order parameter or the size derived from it, take a place close to the lower edge up to the upper edge:
- FIG. 8A shows the correlation of the center thickness of spectacle lenses, which were calculated using the surface model, and the measured center thickness of test spectacle lenses.
- the center thickness (in mm) measured using test data or test spectacle lenses is plotted on the abscissa of FIG. 8A and the center thickness (in mm) calculated according to the model is plotted on the ordinate.
- Figure 8B shows a histogram of the frequency of center thickness residuals (in mm), i.e. the deviations of the calculated from the measured center thickness of test lenses.
- 8C shows the correlation of the back surface curvature of the first principal meridian (curvature of the first principal meridian of the back surface or back surface curvature 1) of spectacle lenses, which were calculated using the surface model, and the measured back surface curvature of the first principal meridian of test spectacle lenses.
- the back surface curvature 1 (in dpt) measured using test data or using test spectacle lenses is plotted on the abscissa of FIG. 8C.
- On the ordinate is the back surface curvature 1 (in dpt) calculated according to the surface model applied.
- 8D shows a histogram of the frequency of the residuals of the back surface curvature 1 (in dpt), ie the deviations of the calculated from the measured back surface curvature 1.
- FIG. 8E shows the correlation of the back surface curvature of the second principal meridian (curvature of the second principal meridian of the back surface or back surface curvature 2) of spectacle lenses, which were calculated using the surface model, and the measured back surface curvature of the first principal meridian of test spectacle lenses.
- the back surface curvature 2 (in dpt) measured using test data or using test spectacle lenses is plotted on the abscissa of FIG. 8E.
- the rear surface curvature 2 (in dpt) calculated according to the surface model is plotted on the ordinate.
- Fig. 8F shows a histogram of the frequency of the residuals of the back surface curvature 2 (in dpt), i.e. the deviations of the calculated from the measured back surface curvature 2.
- 8G shows the correlation of the vertex power in the first principal meridian (vertex power in principal meridian 1) of spectacle lenses, which were calculated using the surface model, and the measured vertex power in the first principal meridian of test spectacle lenses.
- the vertex power measured using test data or test spectacle lenses in principal section 1 (in dpt) is plotted on the abscissa of FIG. 8G and the vertex power in principal section 1 (in dpt) calculated according to the model is plotted on the ordinate.
- 8H shows a histogram of the frequency of deviations (differences) of the vertex power calculated according to the model in principal section 1 from the measured vertex power in principal section 1 (in dpt).
- 8I shows the correlation of the vertex power in the second principal meridian (vertex power in principal meridian 2) of spectacle lenses, which were calculated using the surface model, and the measured vertex power in the second principal meridian of test spectacle lenses.
- the vertex power measured using test data or test spectacle lenses in principal section 2 (in dpt) is plotted on the abscissa of FIG. 8I and the vertex power in principal section 2 (in dpt) calculated according to the model is plotted on the ordinate.
- 8J shows a histogram the frequency of deviations (differences) between the vertex power calculated according to the model in principal section 2 and the measured vertex power in principal section 2 (in dpt).
- the surface model presented here as an example could also be expanded to include the order parameters prism and prism base, e.g. by expanding the surface model to include the prism itself and, as a derived parameter, the angle between the prism base and the axis of astigmatism.
- a surface model designed as a regression model is determined below without having to resort to the data of ophthalmic lenses that have already been calculated. Instead, the parameters of the surface model are formed directly by minimizing the sum of the objective functions, which are used according to a known method from the prior art for (iteratively) calculating a large number of lenses.
- the surface model determined in this way is thus able to carry out the calculation of a single-vision lens with any power, which was specified exclusively via one or more target functions as well as the desired power and diameter.
- An exemplary prior art objective function for a single lens i whose vertex power is characterized by the two principal meridians K1 i , and K2 i , a spherical front surface curvature KVFL i , and a diameter D i owns is as follows:
- the center thickness is given by d M,i , and the order values of the two principal sections are given by and .
- the minimum allowed mean and edge thicknesses are denoted by and . and denote the square deviations of the principal cuts of the vertex power, and the square deviations of the diameter of the lenses, as well as and the square ones Deviations in the center thickness and the edge thicknesses in the two main sections of the back surface.
- the square deviations were calculated as follows, it should be noted that the diameter deviations are calculated from the curvature of the front surface and from the curvature of the back surface in their two principal sections:
- the minimum permitted center and edge thicknesses are constant for all lenses in this example, but they can also be Be functions that can depend on the lens material, diameter, target power or even coating of the lens. These minimum allowable center and edge thickness values represent target values.
- the objective function for optimizing the parameters of the surface model is composed of the sum of the objective functions for individual lenses, summing over all lenses i from the respective data set (i.e. training, validation or test data set), and the center thicknesses and curvatures of the back surface now depend parametrically on the parameters of the surface model:
- ⁇ ( ⁇ dM , ⁇ K1 , ⁇ K2 ) designates the parameters of the surface model, which can be split into three separate parameter sets of the spline coefficients for the center thickness and the two main curvatures of the rear surface.
- the values specified for each lens for front surface curvature, lens diameter, target Values for the principal curvatures of the vertex power, as well as the minimum center and edge thicknesses are denoted overall by .
- the dataset used to train, validate, and test the trained regression model consists of the same sphere and cylinder values as in the previous example (see Figure 7).
- the same spline-based regression model is also used, which, however, initially has a parameter set as the starting point of the optimization that corresponds to lenses with a center thickness of 2mm and back surface curvatures of -5 dpt each (regardless of their front surface curvature) (i.e. only the parameters that correspond to a constant correspond, i.e.
- the Nelder-Mead algorithm was first used here (with 20000 function evaluations), because it is relatively robust and does not require any derivations based on the parameters.
- the BFGS algorithm (Broyden-Fletcher-Goldfarb-Shanno algorithm) was optimized for 200 iterations because it converges faster to a local optimum.
- the gradients for the latter algorithm were calculated numerically, but they could also be given analytically, which would again speed up the optimization. The number of iterations can be accelerated by a more suitable choice of the starting point for the optimization.
- the parameters of an already determined surface model are used when a new surface model is to be determined, whose objective function for individual lenses differs slightly (e.g. in the minimum thicknesses, in the mutual weighting of the terms, or in an additional term) from the objective function for individual lenses Lenses of the first surface model differs.
- Table 7 In the following it is shown on the basis of the ophthalmic lenses from the test data set that the method according to the invention leads to ophthalmic lenses which have almost identical properties in comparison with a method according to the prior art.
- the center thicknesses, the back surface curvatures and the vertex powers of the lenses are plotted against each other or the histograms of the deviations of these values are calculated.
- the minimum center thicknesses and the minimum edge thicknesses are also observed.
- Figures 9A to 9J show the corresponding results.
- FIG. 9A shows the correlation of the center thickness calculated according to the surface model and the center thickness of spectacle lenses according to the prior art (test spectacle lenses).
- the center thickness of spectacle lenses according to the prior art (in mm) is plotted on the abscissa of FIG. 9A and the center thickness (in mm) calculated according to the model is plotted on the ordinate.
- Figure 9B shows a histogram of the frequency of center thickness residuals (in mm), i.e. the deviations of the center thickness calculated according to the surface model from the center thickness of spectacle lenses according to the prior art.
- 9C shows the correlation of the rear surface curvature of the first principal section (rear surface curvature 1) calculated according to the surface model and the rear surface curvature 1 of spectacle lenses according to the prior art (test spectacle lenses).
- the back surface curvature 1 (in dpt) of spectacle lenses according to the prior art is plotted on the abscissa of FIG. 9C and the back surface curvature 1 (in dpt) calculated according to the model is plotted on the ordinate.
- 9D shows a histogram of the frequency of the residuals of the back surface curvature 1 (in Dpt), i.e. the deviations of the back surface curvature 1 calculated according to the model from the back surface curvature 1 of spectacle lenses according to the prior art.
- FIG. 9E shows the correlation of the back surface curvature of the second principal section (back surface curvature 2) calculated according to the surface model and the back surface curvature 2 of spectacle lenses according to the prior art (test spectacle lenses).
- back surface curvature 2 On the abscissa of Fig. 9E is the back surface curvature 2 (in dpt) of spectacle lenses according to the prior art and on the ordinate the rear surface curvature 2 (in dpt) calculated according to the surface model.
- 9F shows a histogram of the frequency of the residuals of the back surface curvature 2 (in Dpt), ie the deviations of the back surface curvature 2 calculated according to the model from the back surface curvature 2 of spectacle lenses according to the prior art.
- FIG. 9G shows the correlation of the vertex power calculated according to the surface model in the first principal meridian (vertex power in principal meridian 1) and the vertex power in the first principal meridian of spectacle lenses according to the prior art (test spectacle lenses).
- the vertex power in principal section 1 (in dpt) of spectacle lenses according to the prior art is plotted on the abscissa of FIG. 9G and the vertex power in principal section 1 (in dpt) calculated according to the surface model is plotted on the ordinate.
- 9H shows a histogram of the frequency of the deviations (differences) of the vertex power in principal meridian 1 calculated according to the model from the vertex power in principal meridian 1 of spectacle lenses according to the prior art (in dpt).
- 9I shows the correlation of the vertex power calculated according to the surface model in the second principal section (vertex power in principal section 2) and the vertex power in the second principal section of spectacle lenses (test spectacle lenses).
- the vertex power in principal section 2 (in dpt) of spectacle lenses according to the prior art is plotted on the abscissa of FIG. 9I and the vertex power in principal section 2 (in dpt) calculated according to the surface model is plotted on the ordinate.
- 9J shows a histogram of the frequency of deviations (differences) of the vertex power in principal meridian 2 calculated according to the model from the vertex power in principal meridian 2 of spectacle lenses according to the prior art (in dpt).
- FIG. 10 illustrates compliance with minimum edge and center thicknesses.
- the center thickness (in mm) is plotted on the abscissa of FIG. 10 and the edge thickness (in mm) on the ordinate.
- the calculation of the center thickness and the two back surface curvatures for any values of sphere sph and cylinder cyl using the surface model is carried out as in the first example.
- Preferred embodiment variants of the invention have been described above using examples. Individual elements of the described embodiment variants are not limited to the respective embodiment variant. Rather, elements of the design variants can be combined with one another as desired, and new design variants can thereby be created. Furthermore, individual features can be modified. Instead of spline functions for defining the surface model, other suitable functions, e.g. polynomial functions, can also be used. The number of model coefficients or model parameters (e.g. spline coefficients) can also be changed. Furthermore, other representations of the area to be calculated, other order parameter sets, target values, target functions and/or optimization methods can be used.
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Ophthalmology & Optometry (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Optics & Photonics (AREA)
- Eyeglasses (AREA)
- Prostheses (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE102020004840.4A DE102020004840A1 (de) | 2020-08-07 | 2020-08-07 | Verbesserte Berechnung ophthalmischer Linsen |
| PCT/EP2021/071710 WO2022029150A1 (de) | 2020-08-07 | 2021-08-04 | Verbesserte berechnung ophthalmischer linsen |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP4193217A1 true EP4193217A1 (de) | 2023-06-14 |
Family
ID=77358257
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP21755448.4A Pending EP4193217A1 (de) | 2020-08-07 | 2021-08-04 | Verbesserte berechnung ophthalmischer linsen |
Country Status (7)
| Country | Link |
|---|---|
| US (1) | US12607874B2 (de) |
| EP (1) | EP4193217A1 (de) |
| JP (1) | JP7596511B2 (de) |
| CN (1) | CN116194824B (de) |
| CL (1) | CL2023000386A1 (de) |
| DE (1) | DE102020004840A1 (de) |
| WO (1) | WO2022029150A1 (de) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102022200462B3 (de) | 2022-01-17 | 2023-03-30 | Rodenstock Gmbh | Computerimplementiertes Verfahren zum Bestimmen eines Fertigungskorrekturmodells für die Herstellung von ophthalmischen Linsen, Speichervorrichtung, Computerprogrammprodukt, Verfahren sowie Vorrichtung |
| US12272018B2 (en) * | 2022-07-15 | 2025-04-08 | The Boeing Company | Modeling system for 3D virtual model |
| EP4325517B1 (de) * | 2022-08-18 | 2024-11-20 | Carl Zeiss Vision International GmbH | Verfahren und vorrichtungen zur durchführung eines sehtestverfahrens an einer person |
| EP4575615A1 (de) * | 2023-12-20 | 2025-06-25 | Carl Zeiss Vision International GmbH | Verfahren zum berechnen einer daten eines brillenglases |
| US12475140B1 (en) | 2024-01-31 | 2025-11-18 | Zoom Communications, Inc. | Synchronizing loosely coupled systems at a third-party system |
| US12585623B1 (en) * | 2024-01-31 | 2026-03-24 | Zoom Communications, Inc. | Synchronizing loosely coupled systems at a source system |
| JP2025141585A (ja) * | 2024-03-15 | 2025-09-29 | ホヤ レンズ タイランド リミテッド | レンズ設計の選択方法、レンズ設計の選択システム、レンズ設計の選択装置 |
Family Cites Families (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4514061A (en) | 1981-04-02 | 1985-04-30 | American Optical Corporation | Progressive power ophthalmic lenses |
| US4861153A (en) | 1986-12-19 | 1989-08-29 | American Optical Corporation | Progressive addition spectacle lens |
| DE4337369A1 (de) | 1993-11-02 | 1995-05-04 | Rodenstock Optik G | Brillenglas mit progressiver Wirkung |
| US5724258A (en) * | 1996-05-09 | 1998-03-03 | Johnson & Johnson Vision Products, Inc. | Neural network analysis for multifocal contact lens design |
| JP3550985B2 (ja) * | 1997-12-02 | 2004-08-04 | 株式会社デンソー | 神経回路網の検証方法,検証装置及び記録媒体 |
| DE19960826A1 (de) | 1999-12-16 | 2001-07-05 | Rodenstock Optik G | Einstärken-Brillenglas mit Vollkorrektion |
| US6655803B1 (en) * | 2000-06-01 | 2003-12-02 | Inray Ltd. | Wavefront method for designing optical elements |
| FR2906621B1 (fr) | 2006-09-28 | 2008-11-28 | Essilor Int | Procede de determination d'une lentille ophtalmique |
| DE102007062929A1 (de) * | 2007-12-28 | 2009-07-02 | Rodenstock Gmbh | Verfahren zur Berechnung und Optimierung eines Brillenglaspaares unter Berücksichtigung binokularer Eigenschaften |
| EP2177943A1 (de) | 2008-10-16 | 2010-04-21 | Essilor International (Compagnie Générale D'Optique) | Bestimmung eines optischen Systems anhand erweiterter Kriterien |
| EP2207118A1 (de) | 2008-12-31 | 2010-07-14 | Essilor International (Compagnie Générale D'Optique) | Verfahren zum Berechnen eines Systems, beispielsweise eines optischen Systems |
| EP2270577A1 (de) | 2009-06-30 | 2011-01-05 | Essilor International (Compagnie Générale D'Optique) | Verfahren und Vorrichtung zur Erzeugung einer Oberfläche einer optischen Linse |
| EP2325617A1 (de) * | 2009-11-18 | 2011-05-25 | ESSILOR INTERNATIONAL (Compagnie Générale d'Optique) | Verfahren zur Bestimmung der binokularen Leistung eines Brillengläserpaares |
| DE102012000390A1 (de) | 2012-01-11 | 2013-07-11 | Rodenstock Gmbh | Brillenglasoptimierung mit individuellem Augenmodell |
| MX347526B (es) | 2012-11-14 | 2017-04-27 | Essilor Int | Metodo para determinar los parametros opticos de una lente oftalmica. |
| WO2015125868A1 (ja) * | 2014-02-20 | 2015-08-27 | 株式会社ニコン | 眼鏡レンズ設計方法、眼鏡レンズ製造方法、眼鏡レンズ、眼鏡レンズ設計システム、眼鏡レンズ設計プログラムおよび記録媒体 |
| US10410118B2 (en) * | 2015-03-13 | 2019-09-10 | Deep Genomics Incorporated | System and method for training neural networks |
| DE102015205721B4 (de) | 2015-03-30 | 2017-01-19 | Rodenstock Gmbh | Verfahren zum Erstellen eines Designs einer Rezeptfläche einer Multifokallinse und Multifokallinse mit einer solchen Rezeptfläche |
| US10018854B2 (en) * | 2016-06-22 | 2018-07-10 | Indizen Optical Technologies of America, LLC | Custom ophthalmic lens design derived from multiple data sources |
| EP3881129B1 (de) * | 2018-11-15 | 2025-08-20 | Essilor International | Verfahren und system zur bestimmung von parametern, die zur herstellung eines optischen artikels verwendet werden, und zugehöriger optischer artikel |
-
2020
- 2020-08-07 DE DE102020004840.4A patent/DE102020004840A1/de active Pending
-
2021
- 2021-08-04 US US18/040,943 patent/US12607874B2/en active Active
- 2021-08-04 WO PCT/EP2021/071710 patent/WO2022029150A1/de not_active Ceased
- 2021-08-04 JP JP2023508061A patent/JP7596511B2/ja active Active
- 2021-08-04 EP EP21755448.4A patent/EP4193217A1/de active Pending
- 2021-08-04 CN CN202180063368.0A patent/CN116194824B/zh active Active
-
2023
- 2023-02-06 CL CL2023000386A patent/CL2023000386A1/es unknown
Also Published As
| Publication number | Publication date |
|---|---|
| US20230296918A1 (en) | 2023-09-21 |
| DE102020004840A1 (de) | 2022-02-10 |
| WO2022029150A1 (de) | 2022-02-10 |
| CL2023000386A1 (es) | 2023-09-29 |
| JP2023537034A (ja) | 2023-08-30 |
| CN116194824A (zh) | 2023-05-30 |
| US12607874B2 (en) | 2026-04-21 |
| JP7596511B2 (ja) | 2024-12-09 |
| CN116194824B (zh) | 2026-04-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP4193217A1 (de) | Verbesserte berechnung ophthalmischer linsen | |
| EP3598214B1 (de) | Gleitsicht-brillenglas mit variablem brechungsindex und verfahren zu dessen entwurf und herstellung | |
| EP2263114B1 (de) | Umskalierung des sollastigmatismus für andere additionen | |
| DE69938027T9 (de) | Torische ophthalmische Linsen | |
| EP2389610B1 (de) | Variables gleitsichtglasdesign | |
| EP2389611B1 (de) | Designsteuerung mittels eines designvielecks | |
| EP2802935B1 (de) | Brillenglasoptimierung mit individuellem augenmodell | |
| EP2767862B1 (de) | Verordnungs- und individualisierungsabhängige anpassung der objektabstandsfunktion an veränderte objektabstände für die nähe und/oder ferne | |
| WO2004086125A1 (de) | Verfahren zum berechnen eines individuellen progressivglases | |
| DE102009005214A1 (de) | Automatische Gleitsichtglasdesignmodifikation | |
| EP3814828B1 (de) | Computerimplementiertes verfahren, datenverarbeitungssystem zum erzeugen eines zieldesigns sowie computerprogramm, speichermedium mit instruktionen zum erzeugen eines zieldesigns, verfahren zum bereitstellen eines brillenglases, speichermedium mit einer numerischen darstellung eines brillenglases und verfahren zum herstellen eines brillenglases | |
| DE102020128958B4 (de) | Verfahren zum Bestimmen eines Brillenglases, Vorrichtung zum Bestimmen von individuellen biometrischen Daten sowie Vorrichtung zum Herstellen eines Brillenglases | |
| DE102017118219B4 (de) | Computerimplementiertes Verfahren zum Erstellen eines Zieldesigns für die Optimierung einer Freiformfläche eines Brillenglases, Computerprogramm, Speichermedium, Computer sowie Vorrichtung | |
| DE102022200462B3 (de) | Computerimplementiertes Verfahren zum Bestimmen eines Fertigungskorrekturmodells für die Herstellung von ophthalmischen Linsen, Speichervorrichtung, Computerprogrammprodukt, Verfahren sowie Vorrichtung | |
| DE102021000451A1 (de) | System und Verfahren fur die Bestelleingangsverarbeitung und Fertigung von Brillengläsern | |
| DE102017118721B4 (de) | Computerimplementiertes Verfahren und Vorrichtung zum Transformieren einer Mehrzahl an ersten Punkten einer ersten Brillenglasfläche in eine Mehrzahl an zweiten Punkten einer zweiten Brillenglasfläche, computerimplementiertes Verfahren und Vorrichtung zum Bestimmen eines Zielbrillenglases mit einer optimierten Brillenglasfläche und Computerprogramm | |
| DE102023125130A1 (de) | Brillenglas für ein Head-Mounted Display, Head-Mounted Display sowie Verfahren zum Entwerfen und Herstellen eines Brillenglases | |
| WO2002065195A2 (de) | B-spline-optimierung von optischen gläsern und insbesondere brillengläsern mit symmetrieebenen | |
| WO2019141386A1 (de) | Gleitsicht-brillenglas mit variablem brechungsindex und verfahren zu dessen entwurf und herstellung |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20230202 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: EXAMINATION IS IN PROGRESS |
|
| 17Q | First examination report despatched |
Effective date: 20250804 |