CN102073759A - Facial form characteristic parameter-based eyeglass configuration control method - Google Patents

Abstract

The invention discloses a facial form characteristic parameter-based eyeglass configuration control method, which comprises the following steps of: 1) establishing a configuration design-oriented eyeglass product structural unit library according to the structural characteristic of an eyeglass product; 2) determining key points of a facial form characteristic by manual marking through a front head image and a side head image of a customer and extracting a facial form characteristic parameter through curve fitting and calculation; 3) converting other acquired demands of the customer into a customer demand matrix on the basis of the facial form characteristic parameter; 4) calculating the similarity between the customer demands and a selectable eyeglass product structural unit in reference to a formula (3); and 5) performing matched searching in a selectable eyeglass product structural unit library according to a result obtained in the step 4) so as to obtain an eyeglass product configuration scheme. A facial form characteristic parameter-based eyeglass configuration control method which can meet individual demands and has high practicability is provided.

Description

A kind of glasses configuration control method based on the shape of face characteristic parameter

Technical field

The present invention relates to a kind of glasses configuration control method.

Background technology

Along with the variation of people's consumption idea and the expansion of product extension, the consumer changes towards diversified, personalized from single sight protection the demand of glasses, and glasses have become a kind of fashion type consumer goods.The glasses kind of customer demand is very many, style changes also very fast.For example, find that by inquiry general each order of glasses manufacturing enterprise in Wenzhou on average comprises 6～10 kinds of eyewear style, the eyewear style of producing in every month is between 20～30 kinds, every amount of money amount is the hundreds of pair at least, and pairs up to ten thousand at most, the market demand present tangible many kinds, become feature in batches.Under this environment, traditional in the past eyeglass design based on customer group, marketing and production model can not satisfy the individual customized requirement that proposes of client far away.Therefore, (Sun Shouqian such as the Sun Shouqian of Zhejiang University, Huang Qi, Sun Lingyun, old glasses adaption system [J] Beijing of just containing customer-oriented feature: computer-aided design (CAD) and graphics journal, 2005,17 (1): 186～188) proposed according to the customer characteristic parameter, from enterprise existing eyewear style storehouse, select suitable glasses to satisfy consumer's individual demand, promptly set up the glasses adaption system.This is a kind of method that satisfies client's individual demand, but also has a problem, and that every pair of glasses that is exactly in the glasses storehouse is not for specific client design, and the glasses that choose according to above method differ and satisfy consumer's the requirement of wearing surely fully.On the other hand, when enterprise carries out personalized design to product, not only to consider client's demand, also will take the design cost of enterprise into account and the response speed in market.

How obtaining client's individual demand rapidly and accurately, and make full use of the design of recombinating fast of enterprise's existing resources, promptly carry out effective product configuration, is the problem that enterprise must face under diversification market environment and the personalized customer demand condition.Aspect the product configuration, existing so far many researchs.Calendar year 2001, proposition products such as MyungS configurations be on the basis of Product Configuration Model with the constraint of configuration rule under product example refinement process (the Myung S that carries out, Han Sknowledge-based parametric design of mechanicalproducts based on configuration design method[J] Expert Systems with Application2001,21 (2): 99～107), setting up rational Product Configuration Model is the key of carrying out effective product configuration.Experts and scholars study from different aspects Product Configuration Model, Ai Xin waits (Ai Xinhao well, Wang Qifu disposes [J] Beijing based on the product rule of growth tree-model: computer integrated manufacturing system, 2003,9 (4): be product structure information 305～308) with the information decomposition in the product layoutprocedure, configuration management information and user's request information, utilize Object-oriented Technique to provide a kind of growth tree allocation models based on the uniform data source, (Liu Xiaobing such as Liu Xiaobing, Yuan Changfeng, Gao Tianyi, Sun Wei, Wang Wanlei is based on the customer-oriented level type of feature Product Configuration Model [J] Beijing: computer integrated manufacturing system, 2003,7 (7): 527～531) proposed to set up the level type Product Configuration Model in customer demand territory-functional domain-behavior territory-domain-feature representation territory, (Liu Mingzhou such as Liu Mingzhou, Ge Maogen, Liu Zhengqiong, Zhang Mingxin, Jiang Zengqiang based on the constraint customizable products allocation models [J] Beijing: computer-aided design (CAD) and graphics journal, 2006,18 (2): 225～230) the formalization modeling method of employing customizable products, realization is to the structure of customizable products allocation models and customizable products data BOM, (Zhu Yun such as Zhu Yun, Chen Xinzhao, Zhang Li, Generalized Product allocation models research [J] Wuhan that the Zhang Jianjun client drives: Chinese mechanical engineering, 2005,16 (13): 1175～1180) the product allocation problem of producing at the mass customization with modular nature has proposed a kind of Generalized Product allocation models of the expansion virtual module based on client's driving.

Because the type of product and the difference of customer demand, the method for its configuration design also has nothing in common with each other.At present, existing document is primarily aimed at General Theory, method and the technology of product configuration design, and the shape of face characteristic parameter of configuration designing requirement, particularly client that does not have to consider specially eyewear products is to the influence of the personalization configuration design of eyewear products.

Summary of the invention

In order to overcome the deficiency of can't the meeting individual requirements of existing glasses collocation method, poor practicability, the invention provides a kind ofly can meet individual requirements, glasses configuration control method that practicality is good based on the shape of face characteristic parameter.

The technical solution adopted for the present invention to solve the technical problems is:

A kind of glasses configuration control method based on the shape of face characteristic parameter, described glasses configuration control method may further comprise the steps:

1) according to the architectural feature of eyewear products, foundation is towards the eyewear products structural unit storehouse of configuration design, described eyewear products structural unit storehouse comprises lens unit storehouse, mirror holder cell library, mirror pin cell library and footmuff cell library, and described lens unit storehouse, mirror holder cell library, mirror pin cell library and footmuff cell library include in order to data module that characterizes the parts factor and the parametrization driver module in order to style and model are adjusted;

2) by client's positive side head image, adopt manual markings to determine the key point of shape of face feature, extract the shape of face characteristic parameter by curve fitting and calculating;

3), based on the shape of face characteristic parameter, be converted into the customer demand matrix together with other demands of the client who obtains;

4), calculate the similarity between customer demand and the optional eyewear products structural unit, with reference to formula (3):

$S(\mathrm{cp},a)=\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\ω}}_i\mathrm{sim}({\mathrm{cp}}_i,a_i)/\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{\mathrm{\ω}}_i---\left(3\right)$

Wherein, S (cp, a) similarity between expression eyewear products structural unit a and the corresponding customer demand cp, ω _iThe weight of i the attribute of expression structural unit a, sim (cp _i, a _i) expression eyewear products structural unit attribute a _iWith customer demand attribute cp _iBetween similarity; I=1,2 ..., p, p are the quantity of structural unit attribute;

5) according to the result of step 4), in optional eyewear products structural unit storehouse, carry out match search, obtain the eyewear products allocation plan.

As preferred a kind of scheme: in the described step 5), result and two indexs of the significance level of structural unit in eyeglass design according to step 4) are carried out match search in optional eyewear products cell library, in the described structural unit, described significance level ordering is followed successively by eyeglass, mirror holder, mirror pin and footmuff from high to low, and the process of described match search is:

1. starting condition is set, has promptly selected eyewear products structural unit set W=Φ;

2. press the importance degree of eyewear products structural unit in product configuration design to eyewear products structural unit a _jSort, j=1,2 ..., k, k represent the quantity of eyewear products structural unit class;

3. calculate S (cp _j, a _j) value, cp _jBe the customer demand parameter, and press S (cp _j, a _j) from big to small to optional structural unit a _jSort;

4. get the highest, the similarity S (cp of importance degree in the optional eyewear products structural unit _j, a _j) the highest eyewear products structural unit a _j

5. judge a _jWhether corresponding element interferes among the eyewear products structural unit set W with selecting;

6. if do not interfere, then a _jSelected eyewear products structural unit set W in joining, changeed 7., otherwise, select the lower optional eyewear products structural unit of similarity successively, change 5., be empty up to optional eyewear products structural unit, change 4.;

7. concentrate deletion a from optional eyewear products structural unit _j8. class eyewear products structural unit if optional eyewear products structural unit collection is empty, then changes, otherwise, upgrade customer demand parameter c p _j, 3. j=j+1 changes;

8. export eyewear products structural unit set W.

Further, described step 2) in, the process of marker characteristic point is: by obtaining customer demand information, from demand information, isolate the positive side photo of client's head, contents such as the color of comparison film, size and background are done the normalization pre-service then, the then manual place between the eyebrows M (M that defines on standard mesh model _x, M _y, M _z), left ear contact point EL (EL _x, EL _y, EL _z), auris dextra contact point ER (ER _x, ER _y, ER _z), bridge of the nose left side contact point BL (BL _x, BL _y, BL _z), the right contact point BR of the bridge of the nose (BR _x, BR _y, BR _z), eye shape unique point (Q ₁, Q ₂... Q _n).EL and ER, BL and BR should be about x=0 plane symmetry, i.e. EL in theory _x=-ER _x, EL _y=ER _yEL _z=ER _z, BL _x=-BR _x, BL _y=BR _y, BL _z=BR _zThe process of extracting characteristic parameter is: the shape of face of delineating according to unique point and characteristic curve and the shape and the position at each position, extract the value GW=|EL of client's glasses width _x-ER _x|, the value GL=|M of mirror pin length _y-(EL _y+ ER _yThe value BW=|BL of)/2|, nose holder width _x-BR _x|, the value BH=|BL of nose holder height _z+ BR _z-2M _z|, according to shape of glasses unique point (Q ₁, Q ₂... Q _n), utilize the quadratic spline method to generate lens features line F (Q ₁, Q ₂... Q _n).

Further again, in the described step 3), the conversion process of customer demand matrix is: according to the shape of face characteristic parameter that extracts, obtain configuration vector the N=((GW of the personalized configuration of glasses design, F), F, GL, BW, BH), in addition, in product configuration design, client's demand also comprises other qualitative demand, and described other qualitative demands comprise the color of Glasses structure unit, material and lettering, described other qualitative demands are converted into the quantitative values cp of product structure cell attribute by the layering refinement, are expanded the requirement matrix (2) that just obtains client disposing vectorial N:

$\mathrm{CP}=\left[\begin{array}{cccccc}{\mathrm{cp}}_{11}& {\mathrm{cp}}_{12}& \mathrm{\Λ}& {\mathrm{cp}}_{1i}& \mathrm{\Λ}& {\mathrm{cp}}_{1n}\\ {\mathrm{cp}}_{21}& {\mathrm{cp}}_{22}& \mathrm{\Λ}& {\mathrm{cp}}_{2i}& \mathrm{\Λ}& {\mathrm{cp}}_{2n}\\ \mathrm{\Λ}& \mathrm{\Λ}& \mathrm{\Λ}& \mathrm{\Λ}& \mathrm{\Λ}& \mathrm{\Λ}\\ {\mathrm{cp}}_{j1}& {\mathrm{cp}}_{j2}& \mathrm{\Λ}& {\mathrm{cp}}_{\mathrm{ji}}& \mathrm{\Λ}& {\mathrm{cp}}_{\mathrm{jn}}\\ \mathrm{\Λ}& \mathrm{\Λ}& \mathrm{\Λ}& \mathrm{\Λ}& \mathrm{\Λ}& \mathrm{\Λ}\\ {\mathrm{cp}}_{k1}& {\mathrm{cp}}_{k2}& \mathrm{\Λ}& {\mathrm{cp}}_{\mathrm{ki}}& \mathrm{\Λ}& {\mathrm{cp}}_{\mathrm{kn}}\end{array}\right]---\left(2\right)$

Wherein, cp ₁₁=(GW, F), cp ₂₁=F, cp ₃₁=GW, cp ₄₁=BW, cp ₅₁=BH, cp _JiI property value representing j product structure unit class.

Further, in the described step 4), described product structure cell attribute comprises the numeric type attribute, and the calculation of similarity degree process is:

Set x _iBe demand parameter cp _iThrough the value after [0,1] interval mapping, y _iBe product structure cell attribute value a _iThrough the value after [0,1] interval mapping, the similarity between them is calculated with (4):

sim(cp _i，a _i)＝1-|x _i-y _i| (4)

For the attribute with upward property, similarity is calculated with (5):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1-(x_i-y_i)& y_i\≤x_i\\ 1& y_i>x_i\end{array}\right.---\left(5\right)$

For the attribute with downward characteristic, similarity is calculated with (6):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1-(y_i-x_i)& y_i\&GreaterEqual;x_i\\ 1& y_i<x_i\end{array}\right.---\left(6\right).$

Or: in the described step 4), described product structure cell attribute comprises the state type attribute, and the calculation of similarity degree process is:

Set x _iBe demand parameter cp _iState value, y _iBe product structure cell attribute a _iState value, similarity between them is calculated with (7):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1& x_i=y_i\\ 0& x_i\&NotEqual;y_i\end{array}\right.---\left(7\right).$

Again or: in the described step 4), described product structure cell attribute comprises the enumeration type attribute, and the calculation of similarity degree process is:

Set x _iBe demand parameter cp _iThrough the value after [0,1] interval mapping, y _iBe product structure cell attribute value a _iThrough the value after [0,1] interval mapping, the similar usefulness (8) between them is calculated:

sim(cp _i，a _i)＝1-|x _i-y _i| (8)。

Or be: in the described step 4), described product structure cell attribute comprises the geometric type attribute, and the calculation of similarity degree process is:

At first to carry out the conversion of coordinate system, make the demand parameter cp that describes geological information _iWith primary attribute value f _iBe under the same normative reference, calculate the value of similarity again with the Euclidean distance between the figure.

In the described step 4), described Method of Weight Determination is expert's assignment method or client's assignment method.

Technical conceive of the present invention is: at the deficiency of present existence, proposed according to customer demand based on the shape of face characteristic parameter, by product configurator relevant parts of match search in the eyewear products cell library, thereby dispose corresponding eyewear products, and carry out certain correction, to satisfy client's customization requirement.So both satisfy client's individual demand, also considered the factor in enterprise's design cost and market simultaneously.

Make glasses manufacturing enterprise when satisfying client's individual demand well, made full use of the existing product resource of enterprise, thereby reached the purpose that shortens the design cycle, reduces design cost and increase the benefit.

Beneficial effect of the present invention mainly shows: can meet individual requirements, practicality is good.

Description of drawings

Fig. 1 is a design flow diagram of the present invention.

Fig. 2 is the cellular construction of eyewear products.

Fig. 3 is for extracting the positive side image of head of shape of face characteristic parameter.

Fig. 4 is for being the Variant Design 1 of example with the eyeglass.

Fig. 5 is for being the Variant Design 2 of example with the eyeglass

Embodiment:

Below in conjunction with accompanying drawing, the present invention is described further.

With reference to Fig. 1, Fig. 2 and Fig. 3, a kind of glasses configuration control method based on client's shape of face characteristic parameter, described method may further comprise the steps:

1) is foundation with the architectural feature of eyewear products and parts detailed configuration characteristic information at different levels and demand mapping relations, made up eyewear products structural unit storehouse, as the basis of configuration design.

Eyewear products structural unit storehouse comprises lens unit storehouse, mirror holder cell library, pin silk cell library, footmuff cell library etc., and the database of each cell type comprises that mainly data and parametrization drive two parts, and the core of data is structrual descriptions.Be that example illustrates that the structrual description of eyewear products cell library and parametrization drive below with the eyeglass.

1. the data structure of eyeglass

In design, production and the process of eyeglass, the principal element that relates to comprises aspects such as the length, width of the geometric center of boundary curve, camber, eyeglass of eyeglass and boundary curve convex closure rectangle.As shown in Figure 4 and Figure 5, boundary curve has reflected the style and the job operation of glasses; Camber is the curvature of space degree of eyeglass; The geometric center of eyeglass embodies the shape of eyeglass and the relation between the optical centre; The length of convex closure rectangle and the value of width are relevant with the shape of face of joining the mirror people with the model of glasses.The feature of eyeglass adopts signature identification, feature name, position, length, width, outline line and camber to represent, i.e. JP_FEATURE={ID, and Name, (xc, yc), X, Y, Pr, r}, data structure is expressed as follows:

Struct JP_FEATURE{ // lens features

Int ID; // feature label

Char Name[20]; // feature name

Float xc, yc; // feature locations

Float X; // eyeglass length

Float Y; // eyeglass width

Struct PROFILE*Pr; // outline line pointer

Float r; // eyeglass camber

}；

Boundary curve is irregular curve, and that determine curve shape is the some P with feature ₁, P ₂, P ₃P ₄, P ₅, P ₆, P ₇, P ₈, be referred to as key point.The border of eyeglass is a smooth curve by these key points, utilizes the quadratic spline method to carry out match, and effect is fine.Therefore, Pr can be expressed as:

Pr＝G(P ₁，P ₂，P ₃----P _n) (1)

Wherein G represents the quadratic spline method, so the data structure of Pr is:

Struct PROFILE{ // cross-sectional data structure

Int ID; // outline line label

Char Name[20]; // outline line name

Void (* spline) (); // matched curve function pointer

Float*P[2]; // key point array pointer

}；

2. parametrization drive scheme

For the eyeglass on irregular curve border, parametrization drives design generally three kinds of situations:

A, style are constant, and model changes, as shown in Figure 4.Eyeglass a and eyeglass b, boundary curve Pr shape is identical, but variation has taken place in the length of convex closure rectangle and width X, Y;

B, model are constant, and style changes, as shown in Figure 5.Eyeglass a and eyeglass b, the length of convex closure rectangle and width X, Y are constant, but variation has taken place in boundary curve Pr shape;

Variation has all taken place in C, model and style.

Key point P ₁, P ₂, P ₃----P _nTherefore the shape of control curve a, realize by curve a formation curve b, and we can adopt the strategy of " first conversion, regeneration " promptly earlier the key point P of control curve a ₁, P ₂, P ₃----P _nGenerate the key point P of control curve b according to transformation rule ₁ ¹, P ₂ ¹, P ₃ ¹----P _n ¹, again by key point P ₁ ¹, P ₂ ¹, P ₃ ¹----P _n ¹Formation curve b.

2), by two given positive side head images of client, adopt manual markings to determine the key point of shape of face feature, by curve fitting with calculate characteristic parameter at the requirement of glasses personalized design.

1. marker characteristic point

By obtaining customer demand information, from demand information, isolate the positive side photo of client's head, contents such as the color of comparison film, size and background are done the normalization pre-service then, the then manual place between the eyebrows M (M that defines on standard mesh model _x, M _y, M _z), left ear contact point EL (EL _x, EL _y, EL _z), auris dextra contact point ER (ER _x, ER _y, ER _z), bridge of the nose left side contact point BL (BL _x, BL _y, BL _z), the right contact point BR of the bridge of the nose (BR _x, BR _y, BR _z), eye shape unique point (Q ₁, Q ₂... Q _n).EL and ER, BL and BR should be about x=0 plane symmetry, i.e. EL in theory _x=-ER _x, EL _y=ER _yEL _z=ER _z, BL _x=-BR _x, BL _y=BR _y, BL _z=BR _z

2. extract characteristic parameter

The value GW=|EL of client's glasses width is extracted in the shape of face of delineating according to unique point and characteristic curve and the shape and the position at each position _x-ER _x|, the value GL=|M of mirror pin length _y-(EL _y+ ER _yThe value BW=|BL of)/2|, nose holder width _x-BR _x|, the value BH=|BL of nose holder height _z+ BR _z-2M _z|.According to shape of glasses unique point (Q ₁, Q ₂... Q _n), utilize the quadratic spline method to generate lens features line F (Q ₁, Q ₂... Q _n).

3), be converted into the customer demand matrix together with other demands of the client who obtains based on the shape of face characteristic parameter.

According to the shape of face characteristic parameter that extracts, can obtain the personalized configuration of glasses design configuration vector N=((GW, F), F, GL, BW, BH).In addition; in product configuration design; client's demand also comprises other usually and describes qualitatively; for example; the color of Glasses structure unit, material, lettering etc.; in order to make the product configuration design device can carry out effective match search, client's qualitative demand must be converted into the quantitative description of feature in the product structure cell library, set up of the mapping of customer demand territory to the product feature territory.For qualitative demand, can be converted into the quantitative values cp of product structure cell attribute by the layering refinement.Like this, expanded the requirement matrix (2) that just can obtain client to disposing vectorial N:

$\mathrm{CP}=\left[\begin{array}{cccccc}{\mathrm{cp}}_{11}& {\mathrm{cp}}_{12}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{1i}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{1n}\\ {\mathrm{cp}}_{21}& {\mathrm{cp}}_{22}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{2i}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{2n}\\ \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}\\ {\mathrm{cp}}_{j1}& {\mathrm{cp}}_{j2}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{\mathrm{ji}}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{\mathrm{jn}}\\ \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}\\ {\mathrm{cp}}_{k1}& {\mathrm{cp}}_{k2}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{\mathrm{ki}}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{\mathrm{kn}}\end{array}\right]---\left(2\right)$

Wherein, cp ₁₁=(GW, F), cp ₂₁=F, cp ₃₁=GW, cp ₄₁=BW, cp ₅₁=BH, cp _JiI property value representing j product structure unit class.

4) similarity between calculating customer demand and the optional product structure unit

Calculate the similarity between customer demand and the optional eyewear products structural unit, with reference to formula (3):

$S(\mathrm{cp},a)=\underset{i=1}{\overset{p}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_i\mathrm{sim}({\mathrm{cp}}_i,a_i)/\underset{i=1}{\overset{p}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_i---\left(3\right)$

Wherein, S (cp, a) similarity between expression eyewear products structural unit a and the corresponding customer demand cp, ω _iThe weight of i the attribute of expression structural unit a, sim (cp _i, a _i) expression eyewear products structural unit attribute a _iWith customer demand attribute cp _iBetween similarity; I=1,2 ..., p, p are the quantity of structural unit attribute.

1. attribute calculation of similarity degree

Dissimilar attributes, its similarity Calculation Method also has nothing in common with each other, and type commonly used has:

The numeric type attribute: have the attribute of numeric type codomain, for example, diameter, length, angle etc. belong to numerical attribute.

The state type attribute: the codomain of status attribute is the state value more than or equal to 2, without any contact, just belongs to this kind attribute as the material of part between each value.

The enumeration type attribute: the codomain of enumerated attributes is one group of discrete state value, and each state value is with significant sequence permutation, and for example, the color of product is pressed certain series arrangement: blue, green, blue or green, red, brown.

The geometric type attribute: the codomain of geometric attribute is to describe the related data of geological information, and different method for expressing, its property value are also different.For example, if adopt repeatedly parametric equation to describe geometric attribute, its property value is exactly the matrix of coefficients of parameter; If adopt the mode of key measurement point to describe geological information, its property value is exactly the coordinate figure of point.

Except type, the dimension of attribute and span also can produce bigger influence to calculation of similarity degree, and the attribute value that has is very big, the value that has is very little, therefore, and before calculating similarity, must make it to be mapped to [0,1] interval as standardization processing to the value of attribute.The type of parameter and property value and span according to demand, their similarity can be calculated and be divided into 4 classes:

A, numeric type x _iBe demand parameter cp _iThrough the value after [0,1] interval mapping, y _iBe product structure cell attribute value a _iThrough the value after [0,1] interval mapping, the similarity between them can be calculated with (4):

sim(cp _i，a _i)＝1-|x _i-y _i| (4)

For the attribute with upward property (dominant eigenvalue is big more, can meet customer need more), similarity is calculated with (5):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1-(x_i-y_i)& y_i\&le;x_i\\ 1& y_i>x_i\end{array}\right.---\left(5\right)$

For the attribute with downward characteristic (dominant eigenvalue is more little, can meet customer need more), similarity is calculated with (6):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1-(y_i-x_i)& y_i\&GreaterEqual;x_i\\ 1& y_i<x_i\end{array}\right.---\left(6\right)$

B, state type x _iBe demand parameter cp _iState value, y _iBe product structure cell attribute a _iState value, the similarity between them can be calculated with (7):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1& x_i=y_i\\ 0& x_i\&NotEqual;y_i\end{array}\right.---\left(7\right)$

C, enumeration type x _iBe demand parameter cp _iThrough the value after [0,1] interval mapping, y _iBe product structure cell attribute value a _iThrough the value after [0,1] interval mapping, the similarity between them can be calculated with (8):

sim(cp _i，a _i)＝1-|x _i-y _i| (8)

D, geometric type are used for describing the demand parameter cp of geological information _iWith primary attribute value f _iMay adopt different method for expressing, therefore, when calculating similarity, at first will carry out the conversion of coordinate system, they are under the same normative reference, calculate the value of similarity again with the Euclidean distance between the figure.

2. Weight Determination

Weight Determination mainly contains two kinds of methods:

A, expert's assignment method: according to client's demand, utilize expert or individual knowledge or experience, relation between analytic product function and the structure attribute, by relatively setting up judgment matrix between per two attributes, utilization analytical hierarchy process and fuzzy judgment matrix are determined the weight of each attribute.

B, client's assignment method: in customer demand, directly comprise the weight data, determine weight by customer demand.For example, in client's demand, directly certain attribute to product or parts claims.

5) in optional eyewear products cell library, carry out match search according to result and two indexs of the significance level of structural unit in eyeglass design of step 4), obtain eyewear products configuration design proposal.

Utilize eyewear products structural unit importance degree and these two indexs of the similarity between customer demand and the product structure unit in the glasses assembling in territory, optional product structure unit, to carry out match search, and a Glasses structure unit that searches is assembled into whole pair of glasses (according to actual needs the structural unit that searches being made Variant Design in the assembling process), obtain glasses configuration design proposal, specific algorithm is as follows:

1. starting condition is set, has promptly selected eyewear products structural unit set W=Φ;

2. press the importance degree of eyewear products structural unit in product configuration design to eyewear products structural unit a _jSort, j=1,2 ..., k, k represent the quantity of eyewear products structural unit class;

3. calculate S (cp _j, a _j) value, cp _jBe the customer demand parameter, and press S (cp _j, a _j) from big to small to optional structural unit a _jSort;

4. get the highest, the similarity S (cp of importance degree in the optional eyewear products structural unit _j, a _j) the highest eyewear products structural unit a _j

5. judge a _jWhether corresponding element interferes among the eyewear products structural unit set W with selecting;

6. if do not interfere, then a _jSelected eyewear products structural unit set W in joining, changeed 7., otherwise, select the lower optional eyewear products structural unit of similarity successively, change 5., be empty up to optional eyewear products structural unit, change 4.;

7. concentrate deletion a from optional eyewear products structural unit _j8. class eyewear products structural unit if optional eyewear products structural unit collection is empty, then changes, otherwise, upgrade customer demand parameter c p _j, 3. j=j+1 changes;

8. export eyewear products structural unit set W.

Claims (9)

1. glasses configuration control method based on the shape of face characteristic parameter, it is characterized in that: described glasses configuration control method may further comprise the steps:

1) according to the architectural feature of eyewear products, foundation is towards the eyewear products structural unit storehouse of configuration design, described eyewear products structural unit storehouse comprises lens unit storehouse, mirror holder cell library, mirror pin cell library and footmuff cell library, and described lens unit storehouse, mirror holder cell library, mirror pin cell library and footmuff cell library include in order to data module that characterizes the parts factor and the parametrization driver module in order to style and model are adjusted;

2) by client's positive side head image, adopt manual markings to determine the key point of shape of face feature, extract the shape of face characteristic parameter by curve fitting and calculating;

3), be converted into the customer demand matrix together with other demands of the client who obtains based on the shape of face characteristic parameter;

4) calculate similarity between customer demand and the optional eyewear products structural unit, with reference to formula (3):

$S(\mathrm{cp},a)=\underset{i=1}{\overset{p}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_i\mathrm{sim}({\mathrm{cp}}_i,a_i)/\underset{i=1}{\overset{p}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_i---\left(3\right)$

Wherein, S (cp, a) similarity between expression eyewear products structural unit a and the corresponding customer demand cp, ω _iThe weight of i the attribute of expression structural unit a, sim (cp _i, a _i) the attribute a of expression eyewear products structural unit _iWith customer demand attribute cp _iBetween similarity; I=1,2 ..., p, p are the quantity of structural unit attribute;

5) according to the result of step 4), in optional eyewear products structural unit storehouse, carry out match search, obtain the eyewear products allocation plan.

2. a kind of glasses configuration control method as claimed in claim 1 based on the shape of face characteristic parameter, it is characterized in that: in the described step 5), result and two indexs of the significance level of structural unit in eyeglass design according to step 4) are carried out match search in optional eyewear products cell library, in the described structural unit, described significance level ordering is followed successively by eyeglass, mirror holder, mirror pin and footmuff from high to low, and the process of described match search is:

1. starting condition is set, has promptly selected eyewear products structural unit set W=Φ;

2. press the importance degree of eyewear products structural unit in product configuration design to eyewear products structural unit a _jSort, j=1,2 ..., k, k represent the quantity of eyewear products structural unit class;

3. calculate S (cp _j, a _j) value, cp _jBe the customer demand parameter, and press S (cp _j, a _j) from big to small to optional structural unit a _jSort;

4. get the highest, the similarity S (cp of importance degree in the optional eyewear products structural unit _j, a _j) the highest eyewear products structural unit a _j

5. judge a _jWhether corresponding element interferes among the eyewear products structural unit set W with selecting;

6. if do not interfere, then a _jSelected eyewear products structural unit set W in joining, changeed 7., otherwise, select the lower optional eyewear products structural unit of similarity successively, change 5., be empty up to optional eyewear products structural unit, change 4.;

7. concentrate deletion a from optional eyewear products structural unit _j8. class eyewear products structural unit if optional eyewear products structural unit collection is empty, then changes, otherwise, upgrade customer demand parameter c p _j, 3. j=j+1 changes;

8. export eyewear products structural unit set W.

3. the glasses configuration control method based on the shape of face characteristic parameter as claimed in claim 1 or 2, it is characterized in that: described step 2), the process of marker characteristic point is: by obtaining customer demand information, from demand information, isolate the positive side photo of client's head, contents such as the color of comparison film, size and background are done the normalization pre-service then, the then manual place between the eyebrows M (M that defines on standard mesh model _x, M _y, M _z), left ear contact point EL (EL _x, EL _y, EL _z), auris dextra contact point ER (ER _x, ER _y, ER _z), bridge of the nose left side contact point BL (BL _x, BL _y, BL _z), the right contact point BR of the bridge of the nose (BR _x, BR _y, BR _z), eye shape unique point (Q ₁, Q ₂... Q _n).EL and ER, BL and BR should be about x=0 plane symmetry, i.e. EL in theory _x=-ER _x, EL _y=ER _yEL _z=ER _z, BL _x=-BR _x, BL _y=BR _y, BL _z=BR _zThe process of extracting characteristic parameter is: the shape of face of delineating according to unique point and characteristic curve and the shape and the position at each position, extract the value GW=|EL of client's glasses width _x-ER _x|, the value GL=|M of mirror pin length _y-(EL _y+ ER _yThe value BW=|BL of)/2|, nose holder width _x-BR _x|, the value BH=|BL of nose holder height _z+ BR _z-2M _z|, according to shape of glasses unique point (Q ₁, Q ₂... Q _n), utilize the quadratic spline method to generate lens features line F (Q ₁, Q ₂... Q _n).

4. the glasses configuration control method based on the shape of face characteristic parameter as claimed in claim 3, it is characterized in that: in the described step 3), the conversion process of customer demand matrix is: according to the shape of face characteristic parameter that extracts, obtain configuration vector the N=((GW of the personalized configuration of glasses design, F), F, GL, BW, BH), in addition, in product configuration design, client's demand also comprises other qualitative demand, and described other qualitative demands comprise the color of Glasses structure unit, material and lettering, described other qualitative demands are converted into the quantitative values cp of product structure cell attribute by the layering refinement, are expanded the requirement matrix (2) that just obtains client disposing vectorial N:

$\mathrm{CP}=\left[\begin{array}{cccccc}{\mathrm{cp}}_{11}& {\mathrm{cp}}_{12}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{1i}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{1n}\\ {\mathrm{cp}}_{21}& {\mathrm{cp}}_{22}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{2i}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{2n}\\ \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}\\ {\mathrm{cp}}_{j1}& {\mathrm{cp}}_{j2}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{\mathrm{ji}}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{\mathrm{jn}}\\ \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}& \mathrm{\&Lambda;}\\ {\mathrm{cp}}_{k1}& {\mathrm{cp}}_{k2}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{\mathrm{ki}}& \mathrm{\&Lambda;}& {\mathrm{cp}}_{\mathrm{kn}}\end{array}\right]---\left(2\right)$

Wherein, cp ₁₁=(GW, F), cp ₂₁=F, cp ₃₁=GW, cp ₄₁=BW, cp ₅₁=BH, cp _JiI property value representing j product structure unit class.

5. the glasses configuration control method based on the shape of face characteristic parameter as claimed in claim 1 or 2 is characterized in that: in the described step 4), described product structure cell attribute comprises the numeric type attribute, and the calculation of similarity degree process is:

Set x _iBe demand parameter cp _iThrough the value after [0,1] interval mapping, y _iBe product structure cell attribute value a _iThrough the value after [0,1] interval mapping, the similarity between them is calculated with (4):

sim(cp _i，a _i)＝1-|x _i-y _i| (4)

For the attribute with upward property, similarity is calculated with (5):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1-(x_i-y_i)& y_i\&le;x_i\\ 1& y_i>x_i\end{array}\right.---\left(5\right)$

For the attribute with downward characteristic, similarity is calculated with (6):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1-(y_i-x_i)& y_i\&GreaterEqual;x_i\\ & \end{array}\right.---\left(6\right).$

$\begin{array}{cc}1& y_i<x_i\end{array}$

6. the glasses configuration control method based on the shape of face characteristic parameter as claimed in claim 1 or 2 is characterized in that: in the described step 4), described product structure cell attribute comprises the state type attribute, and the calculation of similarity degree process is:

Set x _iBe demand parameter cp _iState value, y _iBe product structure cell attribute a _iState value, similarity between them is calculated with (7):

$\mathrm{sim}({\mathrm{cp}}_i,a_i)=\left\{\begin{array}{cc}1& x_i=y_i\\ 0& x_i\&NotEqual;y_i\end{array}\right.---\left(7\right).$

7. the glasses configuration control method based on the shape of face characteristic parameter as claimed in claim 1 or 2 is characterized in that: in the described step 4), described product structure cell attribute comprises the enumeration type attribute, and the calculation of similarity degree process is:

Set x _iBe demand parameter cp _iThrough the value after [0,1] interval mapping, y _iBe product structure cell attribute value a _iThrough the value after [0,1] interval mapping, the similar usefulness (8) between them is calculated:

sim(cp _i，a _i)＝1-|x _i-y _i| (8)。

8. the glasses configuration control method based on the shape of face characteristic parameter as claimed in claim 1 or 2 is characterized in that: in the described step 4), described product structure cell attribute comprises the geometric type attribute, and the calculation of similarity degree process is:

At first to carry out the conversion of coordinate system, make the demand parameter cp that describes geological information _iWith primary attribute value f _iBe under the same normative reference, calculate the value of similarity again with the Euclidean distance between the figure.

9. the glasses configuration control method based on the shape of face characteristic parameter as claimed in claim 1 or 2 is characterized in that: in the described step 4), described Method of Weight Determination is expert's assignment method or client's assignment method.

Images