EP3186767A1

EP3186767A1 - Recommendation system, method, computer program and corresponding medium

Info

Publication number: EP3186767A1
Application number: EP15759916.8A
Authority: EP
Inventors: Philippe WARGNIER
Original assignee: Np2w
Current assignee: Np2w
Priority date: 2014-08-25
Filing date: 2015-08-21
Publication date: 2017-07-05
Also published as: FR3025044B1; FR3025044A1; WO2016030620A1

Abstract

The invention relates to a recommendation system for selecting a spectacle frame, including a database, said data including a digital model of each frame, and a computer apparatus connected to the database and including a camera device, a display screen and an acquisition interface, said computer apparatus being designed for inviting the user to answer a questionnaire, said computer apparatus including an image analysis tool designed for determining, from images of the user, anatomical features of the user, said computer apparatus allowing the user to virtually try on each frame by implementing an augmented reality process, said computer apparatus including processing means capable of producing a selection of frames based on the responses of the user to said questionnaire and said anatomical features. The invention is useful in the field of spectacles.

Description

RECOMMENDATION SYSTEM, METHOD, COMPUTER PROGRAM AND

SUPPORT CORRESPONDING

TECHNICAL AREA

The present invention relates to the general technical field of eyewear, and more specifically to computerized systems for selecting frames adapted to a given user.

The present invention more particularly relates to a recommendation system for helping a user to choose a spectacle frame from a plurality of frames.

The invention also relates to an automated recommendation method for assisting a user in selecting a spectacle frame from a plurality of frames.

The invention further relates to a computer program comprising computer program code means adapted to perform the steps of an automated recommendation method for assisting a user in choosing a spectacle frame from a plurality of frames when said program is executed. on a computer.

The invention also relates to such a computer program implemented on a computer readable medium, and a support that can be read by a computer and on which is recorded such a program.

PRIOR ART

The choice of a new pair of glasses, and more specifically a new frame, is generally a difficult task for many consumers, given the multiplicity of criteria to be taken into account, some of which are very important. made objective and intangible (size of the frame adapted to the anatomical characteristics of the face of the person concerned, especially as regards the interpupillary distance, ability of the frame to support particularly heavy and / or thick glasses ...) and more subjective criteria related to consumer tastes (which are examples relating to the appearance of the frame and its style) and its budget (economic criteria: price of the mount).

The expertise of an optician proves to be generally indispensable, particularly as regards the anatomical and physiological criteria.

Therefore, despite the development of online commerce, it is still very common for many consumers to buy their new frames, especially when they are intended to wear corrective lenses, from an optical store employing optician-eyewear, not only to benefit from the advice and expertise of the professionals present in the store, but also to be able to try different frames and be able to appreciate in a concrete way their adequacy with the objective and subjective choice criteria to each consumer.

This conventional commercial system, however, has a number of disadvantages.

First of all, if optician-eyewear are, because of their expertise, perfectly competent to guide and advise their customers on the technical (optical correction), it is on the other hand much more difficult to intervene in a relatively objective way in terms of aesthetic criteria, which are essential in terms of choice of frame. There is thus a risk of significant disparity, from one store to another, in helping to choose the frame, the consumer possibly being able, in the most unfavorable cases, to be recommended a mount on the basis of essentially commercial criteria. Secondly, this classic commercial store sales system requires optical stores to have a relatively large stock, which must also be renewed regularly, according to sales flows and changes in fashion. This of course leads to significant constraints, in terms of sales areas, management, and de facto limit the choice of the consumer, who is often obliged to multiply visits to different stores to find a perfectly adequate mount.

In addition, even if a store had a very large choice of frames (for example several hundred or thousands of references), it would be difficult anyway, for the consumer, to be able to make a choice without spending a lot of time on his research, which can also mobilize for a significant time the staff of the store concerned.

The classic sale of frames in stores therefore has a number of disadvantages, including in terms of price. In fact, sales prices are sometimes difficult to identify quickly in the store, and can sometimes be perceived as relatively high, despite the high value-added service rendered, and the different requirements that optical retailers make this service (large stock, large sales area, qualified staff ...).

In an attempt to overcome these disadvantages, it may be possible to use electronic commerce techniques, which would allow consumers, for example via an online sales website, to access a very large range of price mounts. optimal, given the simplification of the supply chain.

However, such eip line sales sites are not necessarily the most appropriate tools to specifically sell glasses, especially given the need to use the expertise of an optician for certain criteria, especially ies. optical criteria (choice of suitable glasses) as well as certain anatomical and physiological criteria.

In addition, faced with a very important choice such as an online sales site, which is not subject to constraints in terms of stock management or exhibition space, the consumer can quickly feel lost and experience great difficulty in choosing a mount that suits him in a plethoric offer.

In order to remedy, at least in part, these drawbacks, it has been proposed an online sales site that proposes to finalize the sale within an optical shop, with the assistance of an optician-eyewear. Such a system thus makes it possible to combine some of the advantages of online sales (in terms of choice and price optimization in particular). and some of the benefits of in-store physical selling (such as advice from eye care opticians, and after-sales services).

However, such an improved online sales device does not solve all the problems mentioned above, especially in terms of choice of the frame,

SUMMARY OF THE INVENTION

The objects assigned to the invention are therefore intended to remedy the various disadvantages listed above and to propose a new system and method of recommendation to help a user to choose optimally (including price) a spectacle frame among a plurality of frames, and which allow in particular each user to quickly identify which frame could suit him from a very large choice of frames and this particularly simple, fast and economical.

Another object of the invention is to propose a new system and method of recommendation to help a user to choose a spectacle frame among a plurality of frames, which can be implemented easily and in all circumstances.

Another object of the invention is to propose a new system and method of recommendation to help a user to choose a spectacle frame among a plurality of frames, which are particularly low in computing resources.

Another object of the invention is to propose a new system and method of recommendation to help a user to choose a spectacle frame from among a plurality of frames, which can be implemented using simple and robust technical means. .

Another object of the invention is to propose a new system and method of recommendation to help a user to choose a spectacle frame among a plurality of frames, the implementation of which is particularly playful and spectacular for the user. Another object of the invention is to propose a new system and method of recommendation to help a user to choose a spectacle frame from a plurality of frames, which provide a particularly reliable, accurate and repeatable result.

Another object of the invention is to propose a new system and method of recommendation for helping a user to choose a spectacle frame from among a plurality of frames, which are particularly suitable for use within a store of convenience. sale.

The objects assigned to the invention are achieved by means of a recommendation system to help a user to choose a spectacle frame from among a plurality of frames, said system comprising on the one hand a database storing data relating to said plurality of frames, said data including at least one three-dimensional digital model of each frame, and secondly a computer apparatus which is connected to the database and which includes at least one camera, a screen of display and an acquisition interface, said computer apparatus being adapted to prompt the user to respond, by means of said acquisition interface, to a questionnaire, said computer apparatus including an image analysis tool adapted to determine automatically, from images of the user obtained through the camera, at least the anatomical characteristics of the user, said app computer hardware further enabling the user to virtually try each mount of his choice by implementing an augmented reality process of superimposing in real time an animated image of the user obtained by means of the setting device views and the three-dimensional digital model of the mount concerned and display the resulting moving image on said screen, said computer apparatus including processing means capable of developing a selection of frames from at least the user's responses auditing said questionnaire and said anatomical features, said computer apparatus being adapted to submit, via the screen, said selection to the user.

The objects assigned to the invention are also achieved by means of an automated recommendation method for assisting a user in selecting a frame from among a plurality of frames, said method comprising connecting a base of data with a computer apparatus, said database storing data relating to said plurality of frames, said data including at least a three-dimensional digital model of each frame, while said computer apparatus includes at least one camera, a screen display, an acquisition interface, an image analysis tool and processing means, said method comprising:

a data entry step, during which said computer apparatus invites the user to answer, by means of said acquisition interface, a questionnaire,

a data acquisition step in the course of which said image analysis tool automatically determines, from user images obtained by the camera, at least the anatomical characteristics of the user; ,

a recommendation step during which said processing means develop a selection of frames from at least the user's answers to said questionnaire and said anatomical characteristics, said selection being submitted, via the screen, to the user,

a virtual fitting step during which said computer apparatus allows the user to virtually try each frame of his choice by implementing an augmented reality process of superimposing in real time an animated image of the user obtained by means of the camera and the three-dimensional digital model of the mount concerned and to display the resulting moving picture on said screen.

The objects assigned to the invention are also achieved by means of a computer program comprising computer program code means adapted to perform the steps of the aforementioned method when said program is executed on a computer.

The objects assigned to the invention are also achieved by means of a computer-readable medium on which the computer program in question is recorded. Other objects and advantages of the invention will appear in more detail on reading the description which follows, which is purely illustrative and not limiting.

BEST MODE OF REALIZING THE INVENTION

According to a first aspect, the invention relates to a recommendation system for helping a user to choose a spectacle frame from among a plurality of frames, and preferably from at least several hundreds, and even several thousand (for example at least 10,000). eyeglass frames. By "eyeglass frames" is meant here conventionally a support on which are intended to be fixed glasses (correctors and / or tinted to provide sun protection) and which is itself intended to rest on the nose and / or ears of a user so as to hold the glasses in front of the eyes of the user.

Conventionally, the frames in question are designed to rest on both the nose (by means of a suitable support) and the ears (not via two branches more or less curved) of the user.

The system according to the invention comprises a database, recorded for example on a computer server, storing data relating to said plurality of frames. The data in question includes at least one three-dimensional digital model of each frame, so that each frame has, in the database, a three-dimensional graphical representation as faithful as possible.

By digital model we mean a virtual piece as such, which, unlike a photograph or an image, has neither background nor background. A numerical model can be manipulated in several dimensions, at least two and preferably three, thus forming a three-dimensional numerical model, and in particular so that a user can visualize each of its shapes, faces, edges, etc., without any effort on his part. Advantageously, the digital model intrinsically contains data relating to the modeled element, such as dimensions, shapes, textures, colors, materials, etc. Advantageously, the digital model is realized by means of a design tool assisted by a computer and / or a scanner connected to a computer that is able, from real measurements, to build the digital model, that is to say the corresponding virtual part, and this in at least two dimensions, and preferably three. As a result, a virtual model requires computer software that allows a user to view it.

Preferably, said data relating to said plurality of frames includes at least one or more of the following information:

dimensional characteristics of each of said frames (including for example in particular the spacing of the branches, the length of the branches, the size of the glasses, which is for example expressed for example in the form of an area);

- shape of each of said frames, that is to say the shape of the supports for accommodating glasses (this form may preferably be part of a predetermined list including for example the following forms: rounded, rectangular, square, to closed contour, open contour, etc.);

color (chosen for example from a predetermined list) of each of said frames, it being understood that it is also possible, for a given mount which has several colors, to record in the database the different colors concerned (main color, secondary color n ° 1, secondary color n ° 2 ...),

- material of each of said frames (part of a list comprising for example: chromed metal, brushed aluminum, steel, carbon, scales, plastic ...),

- commercial description of each of the frames (name of the mark, model name, reference number ...),

- selling price of each of said frames,

- etc.

The above list is of course absolutely not limiting and may include other information, for example of a somewhat more subjective nature, such as data relating to the overall appearance and style of each frame, such as the rate of discretion "of each mount (which can be for example a note between 1 and 3 depending on the nature more or less" discreet "of the mount concerned).

The system according to the invention also comprises a computer apparatus, that is to say a programmable digital electronic machine, of the computer type, comprising means (processor, memory, etc.) enabling it to execute computer programs. The computer apparatus implemented in the context of the invention is connected to the abovementioned database, so as to be able to access the data recorded in the database and to be able to use them in the context of automated processing, in the species under the control of an application executed by the computer device.

Preferably, the computer apparatus is connected to the database via the Internet.

According to the invention, the computer apparatus includes at least one camera, such as a digital camera or a digital camera, capable of capturing images of a real scene in a punctual manner (in the manner of a camera) and / or sequentially (in the manner of a video camera).

The camera thus allows the computer apparatus to obtain a digital image faithfully reflecting the visual appearance of the scene located in the field of view of the camera.

The computer apparatus also comprises a display screen, which is for example in the form of a substantially rectangular slab (of the LCD panel type) and preferably allows a color display.

Preferably, the system comprises a device for adjusting the height positioning of the display screen and / or the camera, in order to position the display screen and / or the device in a precise and stable manner. for example, in such a way that the optical axis of the camera crosses the line of the user's eyes or, failing that, passes through the user's face, whatever the size of the latter. slightly above or below the eye line (eg in a range of ± 15 cm). For example, the height positioning adjustment member comprises a support on which are intended to be arranged the camera and / or the display screen, said support being mounted on one or more actuators actuated by a user to adjust the altitude of the support. Preferably, the camera and the display screen can be mounted on separate cylinders. Advantageously, the cylinder or cylinders are actuated automatically according to orders issued by the recommendation system, and in particular to vary the height of the device to automatically adapt to the different steps of the recommendation process.

Preferably, in order to optimize the shooting, the system comprises a lighting device designed to illuminate the face of the user in front. Said lighting device includes, for example, light sources (for example LED lamps) positioned around the display screen, in a plane substantially parallel to the plane in which the display screen is located, way to directly illuminate the face of the user who is in front of the screen and therefore in front of said light sources. Advantageously, said lighting device allows adjustment of said light sources, especially in terms of colorimetry and intensity. This adjustment can be performed manually or automatically by the system itself, preferably from anatomical features (including skin color and age) of the user whose face is illuminated. In the preferred case where the adjustment is carried out automatically, the anatomical characteristics are determined by an image analysis tool, as described below.

In order to further improve the shooting, it is preferable to reduce or even eliminate the influence of the ambient lighting, which can lead to shadows that could affect the proper functioning of the system. To this end, the system advantageously comprises a light shielding device designed to totally or partially isolate the user's face from the ambient light. Said light shielding device comprises for example a roof intended to be disposed above the head of the user to mitigate or eliminate the influence of possible ambient lighting from the ceiling of the place where the system is installed. The light shielding device can even be advantageously formed by a cabin that can completely isolate the user from the ambient lighting. The computer apparatus also comprises an acquisition interface, that is to say a peripheral device allowing a user to provide instructions and / or information to the computing device, with a view in particular to controlling or controlling the operation of the latter and supply it with data necessary for said operation.

The acquisition interface may take the form of any device known in the field, and may for example comprise a keyboard, a pointing device of the mouse type, etc. Preferably, the display screen is a touch screen which therefore allows, in addition to its conventional display function, a pointing function by touching the screen with one or more fingers of the user or with the tip of the screen. a stylus manipulated by the latter. In this case, the display screen also advantageously forms the acquisition interface.

Preferably, the system according to the invention includes an independent portable computing terminal, such as a touch pad or a smart phone, forming said computing device, the latter being connected to the database via the Internet, by means of a connection with or wireless. In this preferred embodiment, the computer apparatus is in the form of an integrated unit that can be manipulated by hand and that includes at least the camera, the display screen and the interface therein. acquisition. However, it is perfectly conceivable, without departing from the scope of the invention, that the computer apparatus is in the form of a fixed installation (not manually movable) in which the camera, the screen The display and the acquisition interface are not part of a single housing but form respective separate units operatively connected to each other by any appropriate means (cable, wireless link ...).

According to the invention, the aforesaid computer apparatus is adapted to prompt the user to respond, by means of said acquisition interface, to a questionnaire. This questionnaire aims to provide the computer device various information that will be useful for recommending to the user one or more frames among said plurality of frames. For this purpose, the questionnaire aims to indicate preferably to the computer apparatus at least one or more of the following information:

D - identifier of the user (eg first name and / or name and / or insurance number ...); - Particular constraints for the optical correction of the user (need a correction with progressive glasses, need a correction of important level (which can suppose the implementation of glasses particularly thick and heavy which are not necessarily compatible with all frames), need only sunglasses, need only corrective lenses ...);

- possible allergic problems;

- maximum budget amount that the user wants to allocate to the purchase of a mount

- shape of the desired frame (round, oval, rectangular ...),

- desired frame style (classic, modern, extravagant ...),

- dress style of the user ...

Of course, the above list is by no means exhaustive and may include various other criteria that may be considered relevant to assist in the selection and recommendation of frames (eg, dwelling, socio-professional category, recreation). practiced, etc.).

The objective of the questionnaire is to lead the user to express his preferences in a field where the latter will not necessarily think of all the parameters. The questionnaire therefore aims to extract certain preferences of a qualitative nature (for example: a classical model is preferred to an extravagant model) or a quantitative one (budget). The user can also indicate, using the form, psychological characteristics (introverted / extrovert, sensitivity to brands, desire for change ...).

The questionnaire may be in the form of open or closed questions put directly on the screen and to which the user responds sequentially by means of the acquisition interface, or by alternative modes of interaction (the user can for example be offered a virtual assistant, a game, in order to identify his preferences).

It is also conceivable for the computing device to produce "on-the-fly" questions, in order to quickly refine the elicitation process. preferences. Such a question posed "on the fly" preferably has a high discriminating power (example: ask the user to classify three frames in order of preference).

The set of answers provided by the user to the questionnaire contributes to constructing a model of the user, said model being formed of a plurality of characteristics directly entered by the user in response to said questionnaire or automatically deduced from said answers by the user. computer apparatus.

According to the invention, the construction of the user's model does not rely only on a positive user's approach, consisting in answering in a more or less direct and conscious manner to a questionnaire, but also on detection operations and automatic analysis which in particular make it possible to communicate to the computer apparatus anatomical and / or physiological data characterizing the user.

To this end, the computer apparatus includes an image analysis tool designed to automatically determine, from user images obtained through the camera, at least the anatomical characteristics of the user. Thus, the computer apparatus is designed to capture one or more images of the user, and in particular of the face of the latter, possibly from different angles (for example from the front, and possibly also from the profile), and then to submit these images. digital to an automated processing, performed by means of an application executed by the computer apparatus, for extracting from the image data representative of certain anatomical and / or physiological characteristics of the user.

Preferably, the image analysis tool is designed to determine one or more of the following characteristics:

- color of the skin of the user's face,

- color of the user's eyes (more precisely, the tool should preferably be able to determine the dominant color in the user's eyes, based on the following colors: blue, green, brown and black); - shape of the user's face (for example, the tool must be able to determine the overall shape of the user's face from among the following shapes: round, oval, heart, square (marked jaw) ...) ;

- gender of the user;

- the age range to which the user belongs (for example, the tool should be able to determine the age range of the person being photographed among the following age groups: under 0, 10 to 20 years, 20 to 30, 30 to 45, 45 to 60, over 60);

the width of the user's face at the temples (for this purpose, the user will for example be placed at a fixed distance from the camera, for example between 1 m and 1 m, and the tool will be able to provide a distance in millimeters);

pupil distance from the user (for example, the tool will be able to determine the horizontal distance between the two pupils of the user, for this purpose the user will for example be placed at a fixed distance from the camera - between 1 m and 1, 20 m distance for example - the expected value will be in this case a difference in vision called "intermediate" expressed in millimeters, the tool is advantageously designed to turn this pupillary difference into "intermediate" vision in pupillary distance far vision from a predetermined mathematical formula, known as such);

- the hair color of the user (the tool may for example be able to determine the dominant color in the user's hair, based on the following colors: blond, chestnut, brown, red, others (for the hair colored), etc.);

- user's length of hair length (for example, the tool will be able to determine the length of the user's hair on the basis of slices of the following length: ras (less than 1 cm), short, half long (between the ears and the shoulders), long (lower than the shoulders) ...);

- shape of the current mount which the user may already have (in this case, the camera takes two snapshots of the user, one without the frame already available to the user and the other with said mount, the tool can then propose the detection of the general shape of the current frame of the user among for example the following forms: round, oval, rectangular, square ...);

- Color of the current frame that the user may already have (for example, the tool may propose the detection of the color cast of the current frame among the following main colors: black, white, red, yellow, blue, green, orange, purple, pink, brown / tortoiseshell, transparent / translucent, golden, silver, multicolored ...);

- size of the current mount (the tool may for example propose the detection of the height of the current frame on the basis of the following sizes: small, medium, large, each frame size may correspond to a percentage of recovery of the face) .

Of course, the analysis tool is not necessarily designed to automatically determine all the above characteristics. Preferably, the image analysis tool is designed to determine at least, automatically by image analysis, the essential characteristics that are: sex, age, the shape of the face, the width of the face at the temples and the pupillary distance of the user.

Advantageously, in order to meet the various constraints related to the wearing of progressive lenses, and also to improve the reliability of the virtual fitting described below, it is preferable that the system can access, preferably automatically or semi-automatically. automatic positioning of the pupil of the user in the future mount that the latter will choose.

To this end, the system advantageously comprises a measuring frame of predetermined dimensions carrying visual references, and that the user is invited to wear so that the camera takes a picture of the user wearing said frame. measurement. In this case, the image analysis tool is designed to automatically determine, by analysis of the facial image of the user wearing the measuring frame, the position of the pupils of the user relative to said visual cues, said computer apparatus being designed to automatically determine, from said position of the user's pupils relative to said visual cues of the measurement mount, the position of the user's pupils relative to each frame of the plurality of frames recorded in the database.

According to the invention, the computer apparatus includes processing means capable of developing a selection of frames (for example a selection of three to ten frames) from at least the user's answers to said questionnaire and said anatomical characteristics. automatically determined by the image analysis tool.

Preferably, the processing means are in this case formed by an application executed by the computer apparatus and using all the data collected on the one hand by means of the questionnaire and on the other hand by means of the tool. image analysis system (anatomical characteristics of the face of the user) for automatically offering the user a selection of frames selected from the plurality of frames recorded in the database.

The computer apparatus is of course designed to then submit, for example via the display screen, said selection to the user.

According to the invention, the computer apparatus also allows the user to virtually try each frame of his choice, and in this case preferably each frame of the aforementioned selection of frames, by the implementation of an augmented reality process consisting of superimposing in real time (i) an animated image of the user obtained by means of the camera and (ii) the three-dimensional digital model of the frame concerned, and displaying the animated image result on the display screen of the computing device.

The display screen thus plays the role of a mirror augmented reality that allows the user to observe in real time an image of his face shod "virtually" of the chosen frame. Such a virtual mirror thus makes it possible to virtually test each frame in a particularly realistic manner, the three-dimensional digital model of the frame concerned being furthermore perfectly positioned thanks to the various information of anatomical and / or physiological nature collected automatically, by image processing, by the image analysis tool built into the system. Advantageously, in order for the user to have the impression of being faced with a virtual mirror, the system's reactivity time to perform the implementation of the augmented reality process must be less than half a second and preferentially be of the order of a tenth of a second. In other words, the preceding expression "superimpose in real time" designates a process that is carried out particularly rapidly, in a manner virtually invisible to the user, and typically in one-tenth of a second.

Preferably, in order to further improve the efficiency of the virtual fitting, the system comprises a device for simulating the weight of the frame, which is designed to make the user feel the weight of the frame that he is wearing. try virtually. For example, the device for simulating the weight of the frame includes a certain number of templates of different weights, which are for example in the form of transparent frames, or in any case designed to not be visible on the screen during virtual fitting (for this purpose, it is also possible that the augmented reality process is specifically designed to make the image of the template disappear on the screen). During the virtual fitting step, the system is then designed to automatically invite the user to put on one of said particular templates, the weight of which corresponds as closely to the weight of the selected frame, so that not only the user will see on the screen with the mount on the nose, but will feel the weight.

In order to simulate even more faithfully the wearing of the frame during the virtual fitting, the system may also preferably include a device for simulating the outline of the frame (or corresponding lenses when it comes to 'a pierced mount, without circles). For example, the contour simulation device includes a number of templates of different shapes, designed to not be visible on the screen during

_* virtual fitting (for this purpose, it is also possible that the augmented reality process is specifically designed to make the image of the template disappear on the screen).

In a particular embodiment, the user is invited to assemble a nose support (intended to simulate the weight of the frame) and an armature (intended to simulate the circles or glasses) and to wear the resulting template thus obtained. during the virtual fitting) to feel as accurately as possible the presence of the mount during the virtual fitting. The system according to the invention, which is preferably based on a web interface, allows to determine automatically and quickly frames that will likely be suitable for the user not only from a morphological point of view but also against other more personal criteria (psychology, stylistic ...), in preferably implementing the following process:

- initially, the user informs the system with basic information (for example name, special constraints for its optical correction: allergies, strong correction, progressive lenses, solar / optical);

- Then the image analysis tool analyzes an image of the user's face, captured by means of the shooting device equipping the computer apparatus, to collect information which is in particular of a morphological nature;

- then an additional questionnaire, this time essentially psychological and / or stylistic will allow to define more precisely the tastes of the user in a global way;

all the parameters collected at this stage will then be recorded by the system and used by the latter to determine a selection of frames likely to correspond as closely as possible to the profile of the user determined above.

- the selection of frames will be submitted to the user, who can then choose the frames he wants to try virtually, thanks to the process of augmented reality ("virtual mirror") that the computer device is able to implement as exposed above; the user is thus able to visualize the mount in real situation, on his face.

Preferably, in order to improve the quality of the recommendations automatically provided by the system, the processing means are designed to develop, based on at least the user's answers to said questionnaire and said anatomical characteristics, a list of weighted criteria used to develop said selection of frames.

In other words, each criterion used by the processing means to elaborate the selection (whether it is a criterion entered via the questionnaire or a criterion filled in by the automatic image analysis) is preferably weighted, the system advantageously comprising an administration interface 2251

19 allowing an administrator to adjust the weighting of each of the criteria, it being understood that it is also perfectly possible to supplement or replace this possibility of "human" adjustment by the administrator, that the system implement a setting of automatic weighting.

For example, the system may be able to automatically adjust the weight of the criteria dynamically. Thus, depending on the value of a criterion, the weight of another criterion will be increased or decreased. For example, if a user responds "extrovert" to the question "what is your style? "The criterion" form mounts must match the shape of the face "should no longer have weight in the sorting frames (thus" debride "selection). If on the other hand the user answers "classical" to the question "what is your style? The criterion "form mounts must match your facial shape" will have a strong impact in the automatic selection of mounts by the system.

Advantageously, the criteria can be classified into three groups, namely the group of critical criteria, the group of low-weight criteria and the group of learning criteria.

The determining criteria are absolute criteria that will filter the results presented in the selection of frames. These criteria are linked to particular "business" constraints, or to particular "user" constraints. For example, if the user needs a strong correction of progressive lenses, the system will not be able to offer some mounts because the glasses can not be mounted on it. Another example, if the width of the face of the user at the temples is greater than the spacing between branches of a frame, this mount will not be offered by the system, because its size is not suitable for the user and may compress his temples.

The criteria of low weight are criteria that will primarily define the display order of frames of the selection made automatically by the system. These criteria will not be directly "filtering", but will weight a given list of results. These criteria will also make it possible to establish basic rules in selection, these basic rules being advantageously called to evolve automatically, by self-learning, as will be explained in what follows. Thus, the system will associate with each given value of these "low weight" criteria a set of corresponding frames.

Depending on the value entered by the user for the criteria in question, the corresponding frames will therefore have a greater "weight" in the frame selection system.

The weight added to each result may itself depend on other criteria. For example, if the user has an "extravagant" style of dress, the "face shape" criterion will have no weight on the selection of the frame. On the other hand, if the user has a "classic" style, the "face shape" criterion will have a significant weight to select the frames (for example according to the rule "no round frame for a round face").

Advantageously, the weight of the criteria will evolve according to usage statistics, as mentioned in the following, so that after a large number of uses of the system, the statistical data take over the "ground rules".

Finally, certain criteria called "learning criteria" will not be attached to a selection of corresponding frames, and will only bring weight after statistical learning by the frame selection system. The weight of these criteria will be maintained initially, and will evolve according to the number of uses. Thus the statistics will have an increasingly strong weight on the sorting of the frames presented, and will eventually overcome the basic rules defined by the criteria of low weight, as discussed above.

Advantageously, the system is therefore designed to record the way it is used by each user, in order to use this information to develop said selection of frames. Preferably, all interactions of the user with the system will be recorded, to allow accurate statistics on each user in particular and on all users in general. Advantageously, the geographical location from which the user uses the system is also recorded. In this preferred case, the processing means will elaborate said selection of frames from not only the user's answers to the questionnaire and the anatomical characteristics determined automatically by the image analysis tool, but also from, for example, sales statistics {the system is indeed advantageously provided with an electronic commerce interface allowing the user to purchase directly online, by means of the acquisition interface, a copy of the frame he has chosen ), and questionnaire responses from other users, for example to identify general trends in public taste and fashion.

The processing means will also advantageously take into account models of frames rejected by the user, either during the current recommendation session or during a previous recommendation session, as well as possibly frames already selected, or even purchased, during a previous recommendation session performed by the same user using the system.

Thus, as and when the system, and the corresponding computer application, are used by different users, more and more parameters and statistics can be collected and recorded by the system in the form of a database. data, so that the selection system can use this information to adapt the weight of the criteria in the selection algorithm.

Thus, the more the database recording the way in which the system is used by each user will be provided, the less the basic rules referred to above, regarding the "low weight" criteria, will have importance in the way that will be implemented by the users. processing means for stopping a selection. Advantageously, the system will then rely primarily on its own experience (usage and sales statistics), according to a self-learning process, either during the current recommendation session or during a session. previous recommendation.

Thus, advantageously, the system will be able to record in a local or remote memory (on a remote server for example) global data and statistics per user, it being understood that the system is advantageously designed to associate a selection of frames to a user session identifier. This allows the system, for example, to use data relating to any previous purchases by the user to refine the system's recommendations to the user. Advantageously, for a given use session of the system by a given user, the system returns a selection of frames, in the form of a list ranked in order of relevance, based on:

- fixed rules defined upstream (determining criteria);

- variable rules defined upstream (learning criteria, low weight criteria);

- global statistics (sales statistics, user action statistics)

- data specific to the user concerned by the session (morphological data and answers to the questionnaire, particularly in stylistic matters, particular constraints and economic data);

- user-specific statistics (previous commands, system usage statistics).

The invention also relates as such, in another aspect, to an automated method of recommendation to assist a user in selecting a spectacle frame from a plurality of frames, said method being advantageously implemented by means of the system described above. Therefore, the entire description

) above is valid, mutatis mutandis, for the process according to the invention. Said method thus comprises linking a database with a computer apparatus, said database storing data relating to said plurality of frames as previously discussed. Said data includes, as described above, at least one three-dimensional digital model of each frame, while said computer apparatus includes at least one camera, a display screen, an acquisition interface, a tool image analysis and processing means.

The method according to the invention further comprises a data entry step, during which the computer device invites the user to respond, by means of said acquisition interface, to a questionnaire, as explained in what follows. above. The method further comprises a data acquisition step, in which the image analysis tool automatically determines, from user images obtained by the camera (in the form of photographic snapshots or short videos, for example of a duration of two seconds), at least the anatomical characteristics of the user, as described above.

The method also comprises a recommendation step, during which the processing means develop a selection of frames from at least the user's answers to said questionnaire and said anatomical characteristics, said selection being submitted, for example via the screen display, to the user.

Finally, the method according to the invention comprises at least one virtual fitting step, during which the computer apparatus allows the user to virtually try each frame of his choice by implementing a process of augmented reality consisting in superimposing in real time an animated image of the user obtained by means of the camera and the three-dimensional digital model of the frame concerned, and displaying the resulting animated image on said display screen. The method according to the invention, advantageously implemented by means of the system described above, thus makes it possible to implement the following overall process, preferably via a web interface:

- initially, the user provides the computer device, via the acquisition interface, some basic information about him (for example the user's first name and the special constraints in terms of optical correction: allergies , strong correction, progressive lenses, solar / optical ...);

the user is then invited, for example by means of a message broadcast on the display screen, to remove his glasses (if he already wears them) and to take place at the right of a mark placed on the ground facing the purpose of the camera. The camera then takes a snapshot (or records a short video, for example of a duration of 2 seconds);

- Optionally, the user will be invited, by means of a message broadcast for example on the display screen, to put on the measurement frame, so as to allow the camera to take a second shot (or a second short video) of the user wearing the measuring mount; - the outi! analysis of images then analyzes the images (or videos) thus obtained to deduce a certain number of anatomical and / or morphological and / or physiological characteristics of the user, in order to contribute to define a profile (modeling) of this last ;

- then a questionnaire (for example about psychology, style and budget) will be submitted to the user, through the display screen. The user's responses will be collected by the computer device so as to be processed, by the processing means, with the characteristics deduced from the aforementioned image analysis, to develop a selection of frames that may correspond to the user profile thus determined. For this purpose, the processing means will advantageously also use the various available statistical data, as mentioned above.

- The user can then choose from the selection that is offered, the frames that he wants to try virtually, thanks to the augmented reality module described above allowing him to virtually visualize the mount on his face, realistically.

- At the end of this phase of virtual fitting, the user will be able to reject some mounts of the selection that do not suit him, and on the contrary keep some mounts that suit him. These choices will be advantageously stored by the recommendation system, in order to be taken into account during a possible future session.

- Once the frame has been chosen definitively, the user can validate his final choice and then start an order process that will be finalized for example with a seller of the store where the system is used.

Thus, the recommendation systems and methods according to the invention advantageously make it possible to implement the following functionalities:

- Filtering frames on the basis of morphological information and dimensions of recorded frames (it is a question of proposing morphologically appropriate frames, that is to say ones which are neither too small nor too large and which allow a eye position centered on the mount). This is a very strong filtering criterion (decisive). - selection of frames on the basis of a preliminary segmentation: initially, the system determines which mounts correspond to which answers for the questionnaire submitted to the user, what are the "standard" rules of mount proposal to be put in The invention thus makes it possible to automatically provide an "intelligent" selection of frames based on these segmentations and rules, it being understood that for the same level of relevance, the frames will be presented in a random order (relevance is defined by the rules initials (low weight criteria) and learning as mentioned above).

- constant learning (as the system is used, more and more statistics will be collected, whether they are global statistics for all users (or groups of users) or personalized statistics specific to each user, so that the system can automatically review the weight of the rules, as mentioned above).

- weight of configurable criteria (the weight of each criterion, defined at startup, may be modified by the administrator). The weight of the criteria will preferably be stored in a database that can be accessed by an administration interface, for modification purposes. Refreshing weights can be done in a cyclical and non-constant way (daily for example).

- weight of the criteria and rules of dynamic selection (the system and the corresponding method are able to adapt the weight of the sorting criteria in a dynamic manner, as explained above).

Preferably, in order to guide the user at best in the frame selection phase, the system and the method according to the invention implement a segmentation of the frames. Thus, when displaying the selection of frames developed automatically by the processing means, the screen will display for example three types of predefined frames (for example according to the following categories: traditional, modern, bold). In each of the display areas corresponding to this segmentation, the user will have for example a selection of five frames.

Finally, the invention makes it possible to install, on a modest work surface (for example between 8 and 20 square meters) a recommendation system. particularly powerful, which provides an "intelligent" assistance to the choice of a mount within an extremely supplied catalog (for example more than 10 000 frames) while allowing the user to try "virtually" each model of his choice .

The system and method according to the invention thus make it much easier for the consumer to choose new glasses, and this quickly, cheap (no stock, small sales area) and relatively fun. It is thus possible and preferable to install the system within a store (which may or may not be an optical store), it being understood that the system can also be implemented exclusively online.

Finally, the invention relates as such to the application for implementing the method referred to above and intended for this purpose to be performed by the computer apparatus.

The invention therefore also relates to a computer program comprising computer program code means adapted to perform the steps of a method according to the foregoing description, when said program is executed on a computer (the term "computer" is to be taken here in its widest sense, and refers not only to "classic" computers but also devices like desktop, tablet, personal assistant ...).

The invention also relates as such to the above computer program when it is implemented on a computer readable medium, whatever the nature) of this medium (hard disk, CD, DVD, USB key ... ).

Finally, the invention also relates as such to a support that can be read by a computer and on which is recorded said program.

According to an aspect independent of the foregoing, and which constitutes an independent invention as such, the use of the previously mentioned questionnaire and the user's> answers to the questionnaire may be replaced by a device for automatically determining additional parameters specific to the user, so does not involve the need to use a questionnaire in any form, the other elements of the system remaining unchanged. So according to this aspect independent of the invention, instead of an "active" questionnaire, one can recover parameters specific to the user in a "passive" way without action of the specific user.

Thus, this independent invention relates to a recommendation system for helping a user to choose a spectacle frame from among a plurality of frames, said system comprising on the one hand a database storing data relating to said plurality of frames, said data including at least one three-dimensional digital model of each frame, and secondly a computer apparatus which is connected to the database and which includes at least one camera and a display screen, said computer apparatus including a tool image analysis system designed to automatically determine, from user images obtained by the camera, at least the anatomical characteristics of the user, said computer apparatus also including an automatic determination system an additional parameter specific to the user, said computer apparatus further enabling the user to to virtually try each mount of its choice by the implementation of an augmented reality process of superimposing in real time an animated image of the user obtained by means of the camera and the three-dimensional digital model of the frame concerned and displaying the resulting animated image on said screen, said computer apparatus including processing means capable of developing a selection of frames from at least said anatomical characteristics and said additional parameter specific to the user, said apparatus computer being designed to submit, via the screen, said selection to the user.

Advantageously, said additional parameter specific to the user may for example relate to the location from which the user uses the system. Indeed, the geographical indication of the place (continent, country, city, district, etc.) where the user is located can give an interesting information from which the processing means will advantageously make the selection of frames, particularly of a morphological point of view or style. This determination may preferentially be made from data such as the IP address of said computing device, or its geofocusing (by GPS for example). Preferentially, this parameter adds! the user's style is relative to his style, such as the user's style of dress. Indeed, it is quite possible that the clothing style is automatically identifiable, in particular by means of the image analysis tool present in the system, for example using logos, trademarks or other distinctive signs present on the clothes. Advantageously, this automatic identification also applies to a spectacle frame that would already be worn by the user.

Advantageously, it is also possible for the system to determine this additional parameter specific to the user from a connection, in particular thanks to the Internet network, with a personal space of the user, such as, for example, social networks (F & cebook® for example). Indeed, by accessing this personal space of the user, the processing means can advantageously have access to the user's favorite items, such as his favorite music, his favorite fashion items, etc. From tors, these elements can advantageously be used in the development of the selection of frames, especially in terms of style, thus avoiding explicitly asking the user.

According to yet another independent aspect which constitutes a distinct invention as such, it is conceivable to implement an automated recommendation method to assist a user in selecting a spectacle frame from among a plurality of frames, said method comprising in relation to a database with a computing device, said database storing data relating to said plurality of frames, said data including at least a three-dimensional digital model of each frame, whereas said computing device includes at least one device camera, a display screen, an image analysis tool, a device for automatically determining additional parameters specific to the user and processing means, said method comprising:

a data acquisition step in the course of which said image analysis tool automatically determines, from user images obtained by the camera, at least the anatomical characteristics of the user; ,) and during which said automatic determination device determines an additional parameter specific to the user, 2251

29

a recommendation step during which said processing means develop a selection of frames from at least said anatomical characteristics and said additional parameter specific to the user, said selection being submitted, via the screen, to the user,

a virtual fitting step during which said computer apparatus enables the user to virtually try each frame of his choice by implementing an augmented reality process consisting of superimposing in real time an animated image of the user obtained by means of the camera and the three-dimensional digital model of the mount concerned and to display the resulting moving picture on said screen.

POSSIBILITY OF INDUSTRIAL APPLICATION

The invention finds its industrial application in particular in the design, production and operation of a recommendation system to help a user to choose a spectacle frame among a plurality of frames.

Claims

CLAIMS - A recommendation system for helping a user to choose a spectacle frame from among a plurality of frames, said system comprising on the one hand a database storing data relating to said plurality of frames, said data including at least one digital model three-dimensional of each frame, and secondly a computer apparatus which is connected to the database and which includes at least one camera, a display screen and an acquisition interface, said computer apparatus being designed to prompt the user to respond, by means of said acquisition interface, to a questionnaire, said computer apparatus including an image analysis tool designed to automatically determine, from user images obtained through the device, at least the anatomical characteristics of the user, said computer apparatus further allowing the user the user to virtually try each mount of his choice by implementing an augmented reality process of superimposing in real time an animated image of the user obtained by means of the camera and the three-dimensional digital model of the frame concerned and displaying the resulting animated image on said screen, said computer apparatus including processing means capable of developing a selection of frames from at least the user's responses to said questionnaire and said anatomical characteristics, said apparatus computer being designed to submit, via the screen, said selection to the user. - System according to claim 1 characterized in that it includes an independent portable computer terminal, such as touch pad or ordiphone, forming said computer apparatus, the latter being connected to the database via the Internet network. - System according to claim 1 or 2 characterized in that said data include at least one or more of the following characteristics: dimensional characteristics of each of said frames, shape of each of said frames, color of each of said frames, material of each of said frames, designation each of said frames, the selling price of each of said frames. System according to one of Claims 1 to 3, characterized in that the said image analysis tool is designed to determine one or more of the following characteristics: the color of the skin of the user's face, the color of the eyes of the user user, user's face shape, user's gender, age range to which the user belongs, user's temple width at temple level, pupillary gap of user, hair color of user the user, the user's length of hair length, the shape of the current frame that the user may already have, the color of the current frame that the user may already have.

System according to one of Claims 1 to 4, characterized in that it comprises a measuring frame of predetermined dimensions bearing visual references, the image analysis tool being designed to automatically determine the position of the pupils of the user relative to said visual cues, said computer apparatus being adapted to automatically determine, from said position of the user's pupils relative to said visual cues of the measurement mount, the position of the user's pupils relative to each frame of the plurality of frames.

System according to one of Claims 1 to 5, characterized in that the said questionnaire is intended to indicate to the said computer apparatus at least one or more of the following criteria: the user's identifier, the need for a correction with progressive lenses, the need for a high level correction, need only sunscreens, needs only corrective lenses, maximum amount of budget the user wants to allocate to the purchase of a mount, desired frame shape, desired frame style, dress style 'user.

System according to one of claims 1 to 6 characterized in that said processing means are designed to develop, from at least the user answers to said questionnaire and said anatomical characteristics, a list of weighted criteria used to develop said selection of mounts. - System according to claim 7 characterized in that it comprises an administration interface allowing an administrator to adjust the weighting of each of said criteria. - System according to one of claims 1 to 8 characterized in that it is designed to record the way it is used by each user, to use this information to develop said selection of frames, - System according to one claims 1 to 9 characterized in that it comprises a member for adjusting the height of the display screen and / or the camera. -System according to one of claims 1 to 10 characterized in that it comprises a lighting device designed to illuminate the face of the user in front. - System according to one of claims 1 to 1 1 characterized in that it comprises a light shield device designed to totally or partially isolate the user's face from the ambient light. - System according to one of claims 1 to 12 characterized in that it comprises a device for simulating the weight of the frame, which is designed to make the user feel the weight of the frame he is trying to to try virtually An automated recommendation method for assisting a user in selecting a spectacle frame from a plurality of frames, said method comprising relating a database to a computer apparatus, said database storing data relating to said plurality of frames, said data including at least a three-dimensional digital model of each frame, while said computer apparatus includes at least one camera, a display screen, an acquisition interface, an analysis tool of image and processing means, said method comprising:

a data entry step, during which said computer apparatus invites the user to answer, by means of said acquisition interface, a questionnaire, a data acquisition step in the course of which said image analysis tool automatically determines, from user images obtained by the camera, at least the anatomical characteristics of the user; ,

a virtual fitting step during which said computer apparatus enables the user to virtually try each frame of his choice by implementing an augmented reality process consisting of superimposing in real time an animated image of the user obtained by means of the camera and the three-dimensional digital model of the mount concerned and to display the resulting moving picture on said screen. A computer program comprising computer program code means adapted to perform the steps of a method according to claim 14 when said program is executed on a computer. Computer program according to claim 15 implemented on a computer readable medium. - A computer-readable medium on which is recorded a program according to claim 15.