Method for modelling a customised helmet for a user for additive manufacturing

Abstract

The present invention relates to a method for modelling a customised helmet for a user, the method being characterised in that it comprises carrying out, by data processing means (11) of a first server (1), the steps of: (a) obtaining a three-dimensional representation of a surface of the user's skull; (b) selecting a three-dimensional functional layer model of the helmet from among a plurality of predefined three-dimensional functional layer models according to the three-dimensional representation of the surface of the user's skull; (c) constructing a three-dimensional customised comfort layer model of the helmet that fills the space between the surface of the skull and the functional layer for the selected three-dimensional functional layer model; (d) fusing the selected three-dimensional functional layer model and the constructed three-dimensional customised comfort layer model into a comprehensive helmet model, suitable for additive manufacturing.

Description

[0001] Description

[0002] Title of the invention: Method for modeling a personalized helmet for a user for additive manufacturing.

[0003] GENERAL TECHNICAL FIELD

[0004] The present invention relates to the field of protective helmets. More specifically, it relates to a method for modeling a personalized helmet for the additive manufacturing of such a helmet.

[0005] STATE OF THE ART

[0006] Protective helmets are used in many activities, particularly sports (cycling, hockey, skiing, canoeing, etc.) and protect the user's head in the event of a fall or impact.

[0007] They traditionally comprise three main parts:

[0008] - a rigid outer shell that protects against penetrating objects and abrasions;

[0009] - a “functional” layer, generally made of solid foam, which collapses and absorbs the shock; therefore, the helmets are single-use, i.e. they must be changed after each impact;

[0010] - a “comfort” layer, often an interior padding, which holds the helmet in place and ensures good ventilation.

[0011] The challenge is to manage to absorb more and more shocks without sacrificing the comfort and lightness of the helmet.

[0012] In particular, it was proposed in document WO2022187232 to personalize the comfort layer (a padding), by modeling it according to the shape of the user's head and printing it in 3D.

[0013] This solution allows for a particularly comfortable "custom-made" helmet, but it is not practical because you have to fix the personalized lining in your helmet yourself, and it remains improvable in terms of shock absorption.

[0014] The invention improves the situation.

[0015] PRESENTATION OF THE INVENTION

[0016] The present invention therefore relates, according to a first aspect, to a method for modeling a personalized headset for a user, the method being characterized in that it comprises the implementation by data processing means of a first server of steps of:

[0017] (a) Obtaining a three-dimensional representation of a surface of the user's skull;

[0018] (b) Selecting a three-dimensional functional layer model of the helmet from a plurality of predefined three-dimensional functional layer models, based on said three-dimensional representation of the surface of the user's skull;

[0019] (c) Constructing a three-dimensional model of a custom comfort layer of the helmet filling the volume between the surface of the skull and the functional layer for said selected three-dimensional model of functional layer;

[0020] (d) Merging the selected three-dimensional functional layer model and the constructed three-dimensional custom comfort layer model into a global helmet model, printable by additive manufacturing.

[0021] According to advantageous and non-limiting characteristics:

[0022] Said three-dimensional representation of the surface of the user's skull is generated from at least one image of said user's skull, in particular a depth image.

[0023] Said image of said user's skull is acquired by a terminal on the user, step (a) comprising the implementation by the terminal of the generation of said three-dimensional representation of the surface of the user's skull from at least one image of said user's skull, then the transmission of said three-dimensional representation of the surface of the user's skull from the terminal to the first server.

[0024] The generation of said three-dimensional representation of the surface of the user's skull comprises the transformation (a1) of said image into a raw three-dimensional representation, the extraction (a2) of the part of the raw three-dimensional representation relating to the user's skull, and the positioning (a3) ​​of said selected part of the raw three-dimensional representation in space relative to the headset.

[0025] Generating said three-dimensional representation of the user's skull surface further comprises processing the raw three-dimensional representation so as to remove artifacts and / or biometric features so as to anonymize said raw three-dimensional representation.

[0026] Said generation of said three-dimensional representation of the surface of the user's skull comprises the application of at least one artificial intelligence model trained on a training database.

[0027] Each predefined three-dimensional model of said functional layer is associated with a helmet size from an ordered set of sizes, the selected predefined three-dimensional model being that of the smallest size among the sizes compatible with the surface of the user's skull.

[0028] Each predefined three-dimensional functional layer model is associated with a predefined three-dimensional outer infill layer model, said merging of step (d) also being with the three-dimensional outer infill layer model associated with the selected three-dimensional functional layer model.

[0029] The functional layer and / or the comfort layer has a lattice structure.

[0030] The functional layer has a periodic lattice structure selected from a plane-based lattice structure, a beam-based lattice structure and a surface-based lattice structure, in particular an extruded cylinder structure extending substantially along an axis normal to the surface of the user's skull and / or the comfort layer has a periodic beam-based lattice structure, in particular a hexagonal lattice structure extending substantially along planes tangent to the surface of the user's skull.

[0031] The outer filler layer also has a periodic lattice structure selected from a plane-based lattice structure, a beam-based lattice structure and a surface-based lattice structure, in particular an extruded cylinder structure extending substantially along the axis normal to the surface of the user's skull, with less periodicity than in the functional layer.

[0032] According to a second aspect, the invention relates to a method for manufacturing a personalized helmet for a user, comprising implementing the method for modeling said personalized helmet for said user according to the first aspect, then a step (e) of additive manufacturing said overall model of the helmet with a printer.

[0033] According to advantageous and non-limiting characteristics, the method comprises a step (f) of assembling the helmet with an external rigid shell, an occipital tightening assembly and / or a chin tightening assembly.

[0034] According to a third aspect, the invention relates to the personalized helmet for a user, directly obtained by the method according to the second aspect.

[0035] According to a fourth aspect, the invention proposes a server for modeling a personalized helmet for a user for additive manufacturing, characterized in that it comprises data processing means configured to:

[0036] - Obtain a three-dimensional representation of a surface of the user's skull;

[0037] - Selecting a three-dimensional functional layer model of the helmet from a plurality of predefined three-dimensional functional layer models, based on said three-dimensional representation of the surface of the user's skull;

[0038] Constructing a three-dimensional model of a custom comfort layer of the helmet filling the volume between the surface of the skull and the functional layer for said selected three-dimensional model of functional layer;

[0039] Merge the selected three-dimensional model of functional layer and the constructed three-dimensional model of custom comfort layer into a global model of the helmet, printable by additive manufacturing.

[0040] According to a fifth and a sixth aspect, the invention relates to a computer program product comprising code instructions for executing a method according to the first aspect of modeling a personalized helmet for a user for additive manufacturing, and a storage means readable by computer equipment on which is recorded a computer program product comprising code instructions for executing a method according to the first aspect of modeling a personalized helmet for a user for additive manufacturing.

[0041] PRESENTATION OF FIGURES

[0042] Other characteristics and advantages of the present invention will appear on reading the following description of a preferred embodiment. This description will be given with reference to the appended drawings in which:

[0043] [Fig. 1] Figure 1 is a diagram of a system for implementing the method according to the invention;

[0044] [Fig. 2] Figure 2 shows a longitudinal section of a preferred embodiment of the custom helmet; [Fig. 3] Figure 3 shows a first example of a functional layer structure of the custom helmet;

[0045] [Fig. 4a] Figure 4a illustrates a first example of the structure of the upper layer of the helmet filling;

[0046] [Fig. 4b] Figure 4b illustrates a second example of the helmet's top filling layer structure;

[0047] [Fig. 4c] Figure 4c illustrates a third example of a helmet top filling layer structure with a second example of a helmet functional layer structure;

[0048] [Fig. 5] Fig. 5 shows front and rear views of said preferred embodiment of the custom helmet;

[0049] [Fig. 6] Figure 6 is a flowchart illustrating the steps of an embodiment of the method according to the invention.

[0050] DETAILED DESCRIPTION

[0051] Architecture

[0052] The present invention relates to a method of modeling, i.e. preparing a printable three-dimensional model, of a personalized helmet for a user for additive manufacturing (i.e. 3D printing), in a system as shown in Figure 1.

[0053] The helmet is typically made of a material suitable for additive manufacturing, such as thermoplastic polyurethane (TPU), for example Ultrasint®, and printing can use MJF (multijet fusion) technology. Alternatively, Stiff TPU and SLS powder sintering technology can be used, for example.

[0054] A helmet means any personal protective equipment intended to protect the head against the consequences of a head injury and in particular: - a sports helmet (in particular for cycling, skateboarding, rollerblading, hockey, American football, skiing, surfing, motor sports, water sports, climbing, rugby, horse riding, etc.);

[0055] - a work helmet (for construction sites, factories, handling, etc.)

[0056] - a police helmet;

[0057] - etc.

[0058] Heavy helmets, such as firefighter helmets, combat helmets or ballistic helmets, which must protect well beyond a simple impact and require special materials (metal, composites), making them impossible to 3D print, will be excluded.

[0059] It is important to understand that the helmet is currently modeled in its entirety, i.e. all its structural layers, with the possible exception of non-printable parts (i.e. which cannot be made from the said material suitable for additive manufacturing), i.e. at most

[0060] - an optional external rigid shell (if you want to use a different material like ABS);

[0061] - one or more specific accessories, also optional:

[0062] O an occipital tightening assembly (ring known as a “tum-ring” and temples, it being understood that it remains possible to print the temples); and o a chin tightening assembly (chin strap possibly provided with a chin strap and strap attachment hooks, it being understood that it remains possible to print the hooks).

[0063] In other words, the helmet is fully 3D printable, i.e. in a single piece, without assembly of parts or layers, with the possible exception of the aforementioned elements which must be made of specific materials. It is also personalized, i.e. uniquely adapted to a user. It is recalled that it was only known to 3D print a personalized lining and fix it in a non-personalized helmet. It is noted that fully 3D printed helmets were known (see for example application US20220104574), but such helmets had very limited shock absorption performance and were not customizable, or at best via the aforementioned lining.

[0064] The system shown comprises a first server 1 for implementing the method. Advantageously, there is a second server 2 (which is a learning device as we will see), typically remote and connected to the first server 1 by a network 20 such as the internet network, but which can be confused with the first server 1.

[0065] Each server 1, 2 has data processing means 11, 21 (typically a processor) and data storage means 12, 22 (a memory, for example a hard disk). As will be seen, the data processing means 22 of the second server 2 can store a learning database. For the sake of simplification, the learning database is called “learning base” in the remainder of this description.

[0066] The system may further advantageously comprise, connected to the first server 1 directly or via the network 20:

[0067] - A 3D printer (so-called 3D printer) for printing said personalized helmet, as explained for example in accordance with MJF or SLS technology. It is assumed that said 3D printer and / or the server 1 is capable of processing said printable three-dimensional model so as to enable printing, in particular by generating the appropriate commands of the print head of the 3D printer;

[0068] - A terminal 10, in particular having data processing means and optical acquisition means, such as a user's smartphone, but also a fixed acquisition cabin (for example located in a store). As we will see, terminal 10 advantageously has a depth camera (i.e. several cameras allowing stereovision, a LIDAR, a structured light camera, etc.)

[0069] Helmet

[0070] The helmet that can be personalized using this process has 2 or 3 layers, presenting various structures and various functions, but all 3D printable in a single piece (i.e. all together, so as to have an "integral" helmet and not layers assembled afterwards), as seen in figure 2:

[0071] - a “functional” layer in the sense that it allows the helmet to perform its protective function, its purpose being to absorb shocks. It may have a lattice structure (lattice) or a so-called “complex” structure by extrusion. Preferably, said structure is a periodic lattice structure (i.e. with repetition of a so-called cell pattern) chosen from: o a planar-based lattice structure: this structure includes structures such as a honeycomb pattern and other models starting from a 2D pattern stretched according to the 3 edirection. These structures provide excellent stability in the plane, while maintaining minimal weight.

[0072] O a strut-based lattice structure: this structure consists of a series of “rods” connected in various directions to form the units of the lattice. o a surface-based lattice structure: this structure is made up of 2-dimensional sub-manifolds (surfaces of space), and generated from trigonometric equations. The best known are TPMS (Triply Periodic Minimal Surfaces) of which the gyroid is an example. According to a particularly preferred embodiment, the functional layer has a structure of extruded cylinders extending substantially along an axis normal to the surface of the user’s skull (case of a lattice structure based on planes). Said cylinder is preferably a “circular” cylinder, i.e. with a circle-shaped base, as seen in Figure 3, but the base may be elliptical, hexagonal, or any other shape.The thickness of the cylinder wall can be variable in certain areas (around vents, around the helmet using a thickness gradient) in order to meet impact resistance standards. The thickness of the functional layer is typically between 15 and 25 mm, preferably 20 mm, except at the ends of the helmet where it can be lower to adapt to the design of the helmet and the shape of the shell.

[0073] A comfort layer which is the layer in practice customizable having the purpose of filling the space existing between the user's skull and the functional layer in order to ensure a perfect fit of the helmet. It can have a lattice structure making it possible to ensure adequate comfort when wearing the helmet by avoiding the presence of "hard points" and a homogeneous distribution of pressure over the entire skull. Again, any example of a lattice structure proposed for the functional layer can be taken, and in particular a periodic lattice structure based on beams. Preferably, the comfort layer has a hexagonal lattice structure extending substantially along planes tangent to the surface of the user's skull, in particular on two planes (i.e. the elementary pattern of the lattice structure is a hexagon with a single cell repetition in the axis normal to the skull).The thickness of this layer naturally varies depending on the size and geometry of the users' skull. Finally, an optional outer filling layer (especially if there is an external rigid shell) whose purpose is to fill the space between the functional layer and the shell (so as to allow an aerodynamic profile for the latter) while ensuring the transfer of forces to the functional layer during impact. It can again have a lattice structure (trellis) or a so-called "complex" structure by extrusion, and we can also take any example of lattice structure proposed for the functional layer.According to a first embodiment illustrated by Figure 4a, the outer filling layer also has a structure of extruded cylinders (preferably the same as the functional layer) extending substantially along the axis normal to the surface of the user's skull, but with a lesser periodicity, i.e. fewer cylinders, in particular one in two. Note that the cylinders of the outer filling layer may be a direct extension of certain cylinders of the functional layer. According to a second embodiment illustrated by Figure 4b, the outer filling layer also has a periodic lattice structure based on beams with a pattern (again on tangent planes, in particular two) extending the cylinders of the functional layer substantially along the axis normal to the surface of the user's skull (i.e.the elementary pattern of the lattice structure is the base of the cylinder - a circle if it is a circular cylinder as in the figures - with a single cell repetition in the axis normal to the skull). The thickness of this layer can vary between 0 and 30 mm. According to a third very different mode illustrated by figure 4c, the functional layer has a periodic lattice structure based on TMPS type surfaces (called diamond structure) and the outer filling layer also has a periodic lattice structure based on beams, in particular with a hexagonal pattern like the comfort layer. Ventilation channels (in particular three) can pass through the two / three layers, in particular in a substantially longitudinal direction (between the front and the back of the helmet, the channels passing through the 2 / 3 layers on each side) as seen in figure 2, so as to ensure thermal regulation of the helmet.At the ends of each channel there is a vent forming an opening in the structure. There are therefore six vents in total in the case of three channels, see Figure 5 (viewed from the front and rear of the helmet respectively).

[0074] The presence of the air vents generates a discontinuity in the structure of the functional layer, thus causing a mechanical weakness around the vents. In order to reinforce these areas, the thickness of the structures is preferably increased over a certain distance following a gradient, for example from a thickness of 0.8 mm on the walls of the vents to 0.5 mm thick at 1 cm distance from these same surfaces.

[0075] And as explained, possible arrangements in the structures can be provided to allow the installation of possible accessories, as can still be seen in Figure 2: o Two branches with toothed ends for said occipital tightening assembly; and o a recess in the different structures in order to position the chin strap hooks for said chin tightening assembly.

[0076] We have now shown how the present process makes it possible to model such a helmet that is fully 3D printable but personalized, while offering guaranteed performance.

[0077] Process

[0078] With reference to Figure 6, the present method is implemented by the data processing means 11 of the first server 1, and begins with a step (a) of obtaining a three-dimensional representation of a surface of the user's skull.

[0079] The present method thus takes as input a three-dimensional representation of a surface of the user's skull, potentially obtained from a "raw" three-dimensional representation, via a phase of processing the raw three-dimensional representation, as we will see in particular by deleting / selecting parts, and by cleaning the geometry, and provides as output a model of the helmet ready to be printed.

[0080] These three-dimensional representations are typically meshes defined by a set of vertices (in space) connected by facets, particularly triangular ones. These are, for example, files in OBJ format (.obj).

[0081] Said three-dimensional representation of the surface of the user's skull is advantageously generated from at least one image of said user's skull, in particular a depth image (depth map), that is to say an image in which the "value" of each pixel designates the distance from the camera.

[0082] In this respect, step (a) comprises, as typically explained, the acquisition of said at least one image of said user's skull using the terminal 10 connected to said first server 1 (note that in the case of a depth image, by acquisition of the image, we potentially already mean its reconstruction from other data such as ordinary color images), but the image or even directly the three-dimensional representation of the surface of the user's skull can be obtained in any way and for example simply received from the network 20, or received from the means 12 or any database, or even artificially generated or simulated.

[0083] In the case where one starts from at least one image of the skull, the generation of said three-dimensional representation of the surface of the user's skull from at least one image of said user's skull advantageously comprises up to three sub-steps: the transformation (a1) of said image into a raw three-dimensional representation, the extraction (a2) of the part of the raw three-dimensional representation relating to the user's skull, and the positioning (a3) ​​of said selected part of the raw three-dimensional representation in space relative to the headset. Note that the order of sub-steps (a2) and (a3) ​​can be completely reversed.

[0084] These sub-steps and generally the generation of said three-dimensional representation of the surface of the user's skull are well known to those skilled in the art, and reference may be made to the aforementioned application WO20221 87232. Preferably, said generation of said three-dimensional representation of the surface of the user's skull (i.e. one or more of the sub-steps (a1 to a3)) comprises the application of at least one artificial intelligence model trained on a learning database, as will be seen later.

[0085] To be more precise, the transformation (a1) of said image into a raw three-dimensional representation is an action well known to those skilled in the art, a fortiori if the image is a depth image. It is only a matter, in the case where said three-dimensional representation is a mesh, of defining said set of vertices and facets, in particular via the coordinates of said vertices and / or the directions of the normals at said vertices. Artificial intelligence models are known, and in particular neural networks, in particular convolutional neural networks (CNNs) adapted to them.

[0086] The extraction (a2) of the part of the raw three-dimensional representation relating to the user's skull aims to "cut" this representation, by selecting a part of interest to isolate the rest.

[0087] The idea is essentially to remove the uninteresting parts of the said raw three-dimensional representation. Indeed, the input image potentially represents the entire head of the user, with visible parts such as the eyes and the face, and on the other hand we commonly have a background. It is desirable to extract the part relating to the skull in order to imitate the amount of useless information and especially to anonymize the data, in particular to avoid privacy issues and regulatory constraints related to the manipulation of personal data such as the GDPR. An artificial intelligence model such as a segmentation model or a classification model applied to the vertices of the mesh is particularly suitable.

[0088] Finally, the positioning (a3) ​​of said selected part of the raw three-dimensional representation (or the entire three-dimensional representation if this sub-step is implemented before sub-step (a2)) in space relative to the headset consists of positioning the skull correctly in a reference frame defined by the headset, so that it is in particular correctly oriented. This sub-step can be referred to as alignment. Naturally, we expect front-facing photos of the user, but they could be slightly skewed and / or from above, so that they must be corrected.

[0089] For this, we can quite easily recognize a geometric element of the face (potentially anatomical, like the eyebrows or the ears), and position the skull accordingly, or use a learned model again.

[0090] In addition, the generation of said three-dimensional representation of the surface of the user's skull may further comprise (before and / or after sub-steps (a2) and (a3)) the processing of the raw three-dimensional representation so as to remove artifacts and / or biometric features so as to anonymize said raw three-dimensional representation. By biometric feature is meant any anatomical feature which would make it possible to identify the user, in particular the eyes.

[0091] Note that if step (a3) ​​is carried out correctly, it is possible that the user is no longer identifiable, so that the deletion of biometric features remains optional. It can be implemented, for example, by making up / distorting the parts of the face constituting biometric features that will still be preserved (or before selecting the parts of interest).

[0092] As for the removal of artifacts, or "cleaning" of the three-dimensional representation, this can be implemented in any known way. This means an improvement in the overall quality of the mesh, without any significant impact on the information contained.

[0093] A first operation performed (potentially just after sub-step (a1)) can concern the surface of the mesh. We assume that the surface has a set of properties: it must not have holes, not form a volume (we can also conceptualize these two properties as having exactly one border) and must not self-intersect (none of the mesh facets must intersect another mesh facet).

[0094] A second operation performed alternatively or in addition (in particular after sub-step (a2)) can be the deletion of any residual parasitic sub-meshes. This case can occur when part of the mesh has not been deleted while the entire surrounding mesh has been. This also covers the case where residues were present in the raw three-dimensional representation and they were not deleted by the cutting procedure.

[0095] To remove these residues, we determine all the disjoint sub-meshes in space (based on the adjacency between the vertices), then we keep only the largest sub-mesh.

[0096] All or part of the generation of said three-dimensional representation of the surface of the user's skull (and therefore of sub-steps (a1) to (a3) ​​+ any deletions of biometric artifacts / traits) can be implemented by the data processing means 11 of the first server 1, and / or directly by the terminal 10, for example by installing a suitable application.

[0097] Especially :

[0098] - In a first case, step (a) comprises the obtaining by the first server 1 of at least one skull image, then the complete implementation of the generation of said three-dimensional representation of the surface of the user's skull. Note that this involves regulatory constraints in terms of personal data management (for example, deletion of the image once the representation is generated);

[0099] In a second case, the terminal 10 fully implements the generation and then transmits to the first server 1 said three-dimensional representation of the surface of the user's skull;

[0100] In a third intermediate case, the terminal 10 implements the sub-step step (a1) so as to obtain the raw three-dimensional representation, then the deletion of the artifacts and / or the biometric features so as to anonymize said raw three-dimensional representation. The latter can then be transmitted to the first server 1 for processing (sub-steps (a2) and (a3)) without any personal data problems; etc.

[0101] The method then comprises a step (b) of selecting a three-dimensional model of the functional layer of the helmet from a plurality of predefined three-dimensional models of the functional layer, based on said three-dimensional representation of the surface of the user's skull.

[0102] Indeed, it is assumed that several variants of functional layers have been pre-modeled (the corresponding three-dimensional models being placed in a database typically stored by the means 12), each of which is known to have the required level of performance in terms of shock absorption. And the variant most suited to the surface of the user's skull is chosen, in particular the one that is closest without collision with this surface. This is all the easier when a positioning sub-step (a2) has been implemented.

[0103] Preferably, each predefined three-dimensional model of said functional layer is associated with a helmet size from an ordered set of sizes, in particular XS, S, M, L, XL, etc. In this case, the predefined three-dimensional model selected is that of the smallest size among the sizes compatible with the surface of the user's skull, i.e. without collision. According to a preferred embodiment, the circumference of the skull is measured on said representation of the surface of the skull, and each predefined three-dimensional model is associated with a range of compatible circumferences.

[0104] Note that alternatively it is entirely possible to have various predefined three-dimensional model shapes (more or less oval skull, pointed, etc.), the person skilled in the art will be able to define a criterion automatically allowing the selection of the most suitable model (for example the one without collision with the average distance between the functional layer and the skull which is the smallest).

[0105] If it is desired to model a helmet with an outer infill layer, each predefined three-dimensional functional layer model is preferentially associated with a predefined three-dimensional outer infill layer model. Thus, step (b) implicitly comprises selecting a three-dimensional outer infill layer model of the helmet from a plurality of predefined three-dimensional outer infill layer models, as being the one associated with the selected three-dimensional functional layer model.

[0106] Then, in a step (c), the method comprises constructing a three-dimensional model of a personalized comfort layer of the helmet filling the volume between the surface of the skull and the functional layer for said selected three-dimensional model of functional layer.

[0107] The idea is simply to use the surface of the user's skull and the internal surface of the selected functional layer model to identify a "design space", i.e. the useful volume in which a printable structure, in particular a lattice, can be integrated, and add any ventilation channels and / or accessories.

[0108] The three-dimensional model in this design space can be generated by following density criteria and applying a predefined structure pattern.

[0109] Then, in a step (d), the method comprises in an original manner the merging of the selected three-dimensional model of functional layer and the constructed three-dimensional model of personalized comfort layer into a global model of the helmet.

[0110] In the case of an outer filling layer, said merging of step (d) is also with the three-dimensional model of outer filling layer associated with the selected three-dimensional model of functional layer, i.e. we merge:

[0111] - the selected three-dimensional model of functional layer,

[0112] - the three-dimensional model of the outer filling layer associated with the selected three-dimensional model of the functional layer and

[0113] - the three-dimensional model constructed from the personalized comfort layer, in the said overall model of the helmet.

[0114] Merging multiple printable three-dimensional models is known, in particular it involves performing a Boolean operation to group the meshes of the functional and filling layer with that of the personalized comfort layer and to obtain only one file. To allow a solid connection to be obtained, it is desirable that there is an overlapping zone between the two geometries.

[0115] Note that step (d) advantageously includes the modification of said printable global model so as to uniquely identify said user, in particular with a string of characters (name and first name), but this could be a barcode, a drawing, etc.

[0116] The objective is to introduce a “label” on the model, that is to say a content that is either “engraved” (i.e. as if we were removing material - we will understand that this engraving is obtained directly during 3D printing and that it is just a matter of not putting material where we want the engraving) or added as extra thickness, in particular inside the helmet and in particular at the level of the rear lattice structure, facilitating the identification of the helmet once printed. The addition of the identifier is done like fusion, typically by generating a three-dimensional representation of the identifier and again by a Boolean operation (addition or subtraction)

[0117] Impression

[0118] At the end of step (d), the overall printable model is obtained; it can be extracted, stored, or transmitted, in particular to the printer 3.

[0119] In this respect, according to a second aspect, the invention relates to a method for manufacturing a personalized helmet for a user, comprising the implementation of the method for modeling said personalized helmet according to the first aspect (i.e. steps (a) to (d)), then a step (e) of printing by the printer 3 said global model of the helmet. In other words, step (e) is a step of additive manufacturing of said personalized helmet from its model (the printable global model) with the printer 3, see figure 6.

[0120] It is repeated that the said personalized helmet is the single-piece piece resulting from the 3D printing of the said global helmet model.

[0121] More specifically, printer 3 takes said global model as input, and prints it in its entirety, which results in the creation of said personalized helmet for the user.

[0122] Here, it is the 3D printing software that typically takes over and allows the import of the global model file (typically a PFH file), the configuration of the printing parameters and the control of the printer 3.

[0123] Step (e) may include sandblasting the printed part (the helmet) to remove excess powder and / or chemical etching to smooth the surfaces, remove open porosities, a source of crack initiation and thus improve the durability of the helmet with respect to structural breakage.

[0124] The method may then comprise a step (f) of assembling the helmet with any external rigid shell, occipital tightening assembly and / or chin tightening assembly. In this respect, according to a third aspect, the invention relates to the personalized helmet obtained by the method according to the second aspect.

[0125] Learning

[0126] In the case where said generation of said three-dimensional representation of the surface of the user's skull comprises the application of at least one trained artificial intelligence model (i.e. the algorithm(s) used are obtained by machine learning), the method advantageously comprises a prior step (aO) of learning, by the data processing means 21 of the second server 2, said artificial intelligence model on a learning database stored by the storage means 22 of the second server 2

[0127] The learning base stores, for example, pairs of a three-dimensional representation of a surface of the skull of reference individuals and the corresponding raw three-dimensional representation.

[0128] Alternatively or in addition, said artificial intelligence models can be pre-trained on public databases.

[0129] Server

[0130] According to a fourth aspect, the invention relates to the first server 1 for implementing the method according to the first aspect.

[0131] Thus, this first server 1 comprises, as explained, at least data processing means 21 and a memory 22. This typically involves modeling a personalized headset for a user for additive manufacturing, connected to user terminals 10 and / or stores.

[0132] The data processing means 21 are configured to implement steps consisting of:

[0133] - Obtaining a three-dimensional representation of a surface of the user's skull; - Selecting a three-dimensional functional layer model of the headset from a plurality of predefined three-dimensional functional layer models, based on said three-dimensional representation of the surface of the user's skull;

[0134] - Constructing a three-dimensional model of a custom comfort layer of the helmet filling the volume between the surface of the skull and the functional layer for said selected three-dimensional model of functional layer;

[0135] - Merge the selected three-dimensional model of functional layer and the constructed three-dimensional model of custom comfort layer into a global model of the helmet, printable by additive manufacturing.

[0136] According to another aspect, the invention may propose a system comprising said first server 2, as well as at least one terminal 10 connected (via the network 20). Advantageously, said system also comprises the second server 2, connected to the first server 2 still via the network 20.

[0137] The second server 2 comprises data processing means 21 configured to implement learning.

[0138] Computer program product

[0139] According to a fifth and a sixth aspect, the invention relates to a computer program product comprising code instructions for the execution (on the data processing means 21 of the first server 2) of a method according to the first aspect of modeling a personalized headset for a user, as well as storage means readable by computer equipment (for example the data storage means 22 of the first server 2) on which this computer program product is found.

Claims

CLAIMS 1. Method for modeling a personalized helmet for a user, the method being characterized in that it comprises the implementation by data processing means (11) of a first server (1) of steps of: (a) Obtaining a three-dimensional representation of a surface of the user's skull; (b) Selecting a three-dimensional functional layer model of the helmet from a plurality of predefined three-dimensional functional layer models, based on said three-dimensional representation of the surface of the user's skull; (c) Constructing a three-dimensional model of a custom comfort layer of the helmet filling the volume between the surface of the skull and the functional layer for said selected three-dimensional model of functional layer; (d) Merging the selected three-dimensional functional layer model and the constructed three-dimensional custom comfort layer model into a global helmet model, printable by additive manufacturing.

2. Method according to claim 1, wherein said three-dimensional representation of the surface of the user's skull is generated from at least one image of said user's skull, in particular a depth image.

3. Method according to claim 2, wherein said image of said user's skull is acquired by a terminal (10) on the user, the step (a) comprising the implementation by the terminal (1) of the generation of said three-dimensional representation of the surface of the user's skull from at least one image of said user's skull, then the transmission of said three-dimensional representation of the surface of the user's skull from the terminal (10) to the first server (1).

4. Method according to one of claims 2 and 3, wherein the generation of said three-dimensional representation of the surface of the user's skull comprises the transformation (a1) of said image into a raw three-dimensional representation, the extraction (a2) of the part of the raw three-dimensional representation relating to the user's skull, and the positioning (a3) ​​of said selected part of the raw three-dimensional representation in space relative to the headset.

5. The method of claim 4, wherein generating said three-dimensional representation of the user's skull surface further comprises processing the raw three-dimensional representation so as to remove artifacts and / or biometric features so as to anonymize said raw three-dimensional representation.

6. Method according to one of claims 2 to 5, wherein said generation of said three-dimensional representation of the surface of the user's skull comprises the application of at least one artificial intelligence model trained on a training database.

7. Method according to one of claims 1 to 6, in which each predefined three-dimensional model of said functional layer is associated with a helmet size from an ordered set of sizes, the predefined three-dimensional model selected being that of the smallest size from among the sizes compatible with the surface of the user's skull.

8. Method according to one of claims 1 to 7, wherein each predefined three-dimensional model of functional layer is associated with a predefined three-dimensional model of outer filling layer, said merging of step (d) being also with the three-dimensional model of outer filling layer associated with the selected three-dimensional model of functional layer.

9. Method according to one of claims 1 to 8, in which the functional layer and / or the comfort layer has a lattice structure.

10. The method of claim 9, wherein the functional layer has a periodic lattice structure selected from a plane-based lattice structure, a beam-based lattice structure and a surface-based lattice structure, in particular an extruded cylinder structure extending substantially along an axis normal to the surface of the user's skull and / or the comfort layer has a beam-based periodic lattice structure, in particular a hexagonal lattice structure extending substantially along planes tangent to the surface of the user's skull.

11. A method according to claims 8 and 10 in combination, wherein the outer filling layer also has a periodic lattice structure selected from a plane-based lattice structure, a beam-based lattice structure and a surface-based lattice structure, in particular an extruded cylinder structure extending substantially along the axis normal to the surface of the user's skull, with less periodicity than in the functional layer.

12. Method for manufacturing a personalized helmet for a user, comprising implementing the method for modeling said personalized helmet for said user according to one of claims 1 to 11, then a step (e) of additive manufacturing of said overall model of the helmet with a printer (3).

13. The method of claim 12, comprises a step (f) of assembling the helmet with an external rigid shell, an occipital clamping assembly and / or a chin clamping assembly.

14. Personalized helmet for a user, directly obtained by the method according to one of claims 12 and 13.

15. Server (1) for modeling a personalized helmet for a user for additive manufacturing, characterized in that it comprises data processing means (11) configured to: - Obtain a three-dimensional representation of a surface of the user's skull; - Selecting a three-dimensional functional layer model of the helmet from a plurality of predefined three-dimensional functional layer models, based on said three-dimensional representation of the surface of the user's skull; - Constructing a three-dimensional model of a personalized comfort layer of the helmet filling the volume between the surface of the skull and the functional layer for said selected three-dimensional model of functional layer; - Merge the selected three-dimensional model of functional layer and the constructed three-dimensional model of custom comfort layer into a global model of the helmet, printable by additive manufacturing.

16. Computer program product comprising code instructions for executing a method according to one of claims 1 to 11 for modeling a personalized helmet for a user for additive manufacturing, when said program is executed on a computer.

17. Storage means readable by computer equipment on which is recorded a computer program product comprising code instructions for the execution of a method according to one of claims 1 to 11 for modeling a personalized helmet for an additive manufacturing user.

Images