System and method for the construction of a three-dimensional model of a dental prosthesis for at least one user

Description

1 / 19 “SYSTEM AND METHOD FOR CONSTRUCTING A THREE-DIMENSIONAL MODEL OF A DENTAL PROSTHESIS FOR AT LEAST ONE USER” DESCRIPTIVE MEMORY

[0001] The present invention relates to a system and method for constructing a three-dimensional model of a dental prosthesis for at least one user, intended for the field of dentistry.

[0002] The system and method of the invention work by using passive stereo techniques in conjunction with neural networks, which allows the intraoral device used for image capture to only include the camera(s) necessary for such a task and for all data processing to be performed outside the intraoral device, thus being considerably less complex and costly than state-of-the-art solutions.

[0003] In this sense, the system of the invention essentially comprises at least one intraoral device, to be placed in the oral cavity of at least one user; at least one set of cameras placed in at least one intraoral device, to capture at least one stereo image of the oral cavity of at least one user; and at least one processing means, which receives at least one stereo image; wherein the at least one processing means comprises at least one trained neural network, which analyzes at least one stereo image to estimate at least one depth map; and wherein the at least one processing means further comprises at least one localization and mapping block, which sequentially integrates at least one stereo image and at least one depth map into the generated three-dimensional model.

[0004] On the other hand, the method of the invention essentially comprises the steps of: capturing at least one stereo image through at least one set of cameras arranged in at least one intraoral device; receiving at least one stereo image through at least one processing means; analyzing at least one stereo image to estimate by Petition 870250085553, dated 09 / 22 / 2025, pp. 81 / 108 2 / 19 minus one depth map through at least one trained neural network, comprising at least one processing means; and sequentially integrate at least one stereo image and at least one depth map into the three-dimensional model generated through at least one localization and mapping block, comprising at least one processing means.

[0005] Using the system and method of the invention, it is not only possible to improve accessibility to this type of technology for dental patients, thanks to the simplification of the physical device used in these procedures, but also to provide a solution that achieves greater precision than technologies that use traditional modeling techniques. BACKGROUND

[0006] In the field of dentistry and its various specialties, the use of technologies that allow the modeling of patients' dental prostheses is a widespread practice due to the speed and simplicity of the process.

[0007] Most of these solutions involve the use of an intraoral device or scanner that includes at least one camera that uses one of the known principles for three-dimensional modeling, such as confocal microscopy, active stereo or structured light, in combination with an algorithm that usually performs the processing within the same device.

[0008] In this sense, confocal microscopy is the most complex technique in terms of hardware to use and, therefore, the most expensive to manufacture and implement. This technique calculates the depth of the image by varying the focus of a light source and filtering out-of-focus light with a lens. For each focus adjustment, the camera receives light from the surface areas that are in focus. Since the focus of the light is controlled and the geometry of the lens is known, there is a relationship that allows the depth of the focused areas to be calculated.

[0009] On the other hand, scanners that use the active stereo technique capture pairs of images with a stereo camera, while simultaneously projecting a pattern onto the surface. An algorithm then calculates the Petition 870250085553, dated 09 / 22 / 2025, page 82 / 108 3 / 19 matching between both images using the projected pattern. This technique requires placing a pattern projector inside the scanner or intraoral device, which also makes its construction more complex.

[0010] Finally, structured light scanners project a known light pattern onto the surface and capture images with a camera. By observing how the light pattern deforms on the surface, an algorithm infers the topology and depth of that surface. Again, scanners using this technique require additional hardware, making them more expensive to manufacture.

[0011] Furthermore, since most of these solutions incorporate processing within the intraoral scanner, extreme care must be taken regarding potential damage to the equipment, as repairs can be expensive.

[0012] Therefore, there is a growing need not only for a system and method that simplifies the construction of the intraoral scanner or device by using fewer components, as well as by separating the data acquisition and processing processes of the intraoral device, but also one that improves the quality of the three-dimensional modeling, which has a significant margin for improvement in light of the techniques used in the state of the art.

[0013] In the field of patents, there are solutions that focus on devices or systems for three-dimensional modeling of dental prostheses in the dental field. For example, US patent application US20140146142A1 describes a three-dimensional measuring device for measurement in the absence of active or structured light projection, which includes an image capture device and an image data processor. The image capture device is capable of simultaneously, or almost simultaneously, capturing at least two images, one of which is totally or partially included in the other. The included image describes a narrower field than the other and has greater precision than the other.

[0014] In this sense, document US20140146142A1, although it describes a device that has at least one camera, which processes Petition 870250085553, dated 09 / 22 / 2025, pp. 83 / 108 4 / 19 The information outside of said device, specifically on a computer adapted for this purpose, makes no reference to the use of neural networks previously trained with dental models during information processing, only mentioning the use of various algorithms in the different stages of processing. As the processing of document US20140146142A1 is performed on the computer of the expert performing the procedure, the algorithms used cannot be very complex, as this would require a computer with high processing power, which would make the device accessible to only a few people.Therefore, this document does not provide an algorithm that allows for obtaining results superior to those existing in the state of the art, since its processing capacity will be limited to the processing capacity of the user's computer, a situation that does not occur with the present invention, in which, because there is a means of processing external to the intraoral device, which can be hosted in a cloud, it allows the processing capacity to increase considerably by being able to use, for example, neural networks trained with a large number of parameters, which substantially improves the results of the three-dimensional modeling.

[0015] Another example is that disclosed in international patent application WO2021250091A1, which describes a method for the automatic segmentation of a dental arch, comprising acquiring a three-dimensional surface of the dental arch in order to obtain a three-dimensional representation comprising a set of vertices, generating virtual views from the three-dimensional representation, projecting the three-dimensional representation onto each two-dimensional virtual view to obtain an image representing each vertex of the virtual view, processing each image using a deep learning network, performing an inverse projection of each image to assign to each vertex of the three-dimensional representation one or more pixels from the images in which the vertex appears and to which it corresponds, and assigning one or more probability vectors to each vertex, determining the class of dental tissue to which each vertex belongs.

[0016] When comparing the document description Petition 870250085553, dated 09 / 22 / 2025, pp. 84 / 108 5 / 19 Regarding the present application WO2021250091A1, it is possible to observe that, although the document describes an intraoral device that uses trained neural network techniques for image processing, this algorithm is housed within a processing module located in the intraoral device, which considerably limits the device's processing capacity and consequently limits the complexity of the neural network used. This does not occur in the present invention, where all processing is performed outside the intraoral device, thus not limiting processing capacity and improving the quality of the three-dimensional modeling, allowing the reconstruction of the three-dimensional model to be obtained practically in real time. Furthermore, the construction of this intraoral device is simplified, as it only has the camera(s) necessary to obtain the images and the associated circuitry.

[0017] As can be seen from the documents described above, the vast majority of these devices and systems do not aim to simplify the construction of intraoral devices, with the consequent economic savings, but rather to solve other types of problems, such as avoiding the use of structured light, for example.Although solutions exist that perform data processing outside the intraoral device or that use neural networks for image analysis, these solutions are far from what is described by the present invention, which, in addition to providing a device for simplified construction and use, allows processing to be performed on a server located in the cloud. This not only eliminates the need for the user to have processing-capable equipment on-site, but also significantly improves the quality of the three-dimensional modeling of the dental prosthesis. This is because it is possible to use a much more complex processing algorithm, in combination with neural networks trained with a high number of parameters, to analyze the received images in order to determine the corresponding depth map.

[0018] Therefore, it is necessary to have a system and a method not only to provide a simpler and more economical intraoral device, but also to improve the quality and speed of three-dimensional modeling of Petition 870250085553, dated 09 / 22 / 2025, pages 85 / 108 6 / 19 patient's dental prosthesis, so that the dentist responsible for the procedure can obtain the results in real time. Furthermore, there is a need for a solution that avoids the use of complex and expensive data processing equipment, which the present invention makes possible through a server that receives the information obtained by the intraoral device directly from it or through a computer or other similar electronic device that acts as a link and displays the results. This and other advantages associated with other aspects of the technology are described in more detail below. DESCRIPTION OF THE INVENTION

[0019] The invention relates to a system and method for constructing a three-dimensional model of a dental prosthesis for at least one user in real time, which simplifies the construction of an intraoral device to be used, while improving the quality and speed of three-dimensional modeling.

[0020] According to a first preferred embodiment of the invention, the system for constructing a three-dimensional model of a dental prosthesis for at least one user comprises: [0021 ] - at least one intraoral device, to be placed in the oral cavity of at least one user;

[0022] - at least one set of cameras arranged in at least one intraoral device, to capture at least one stereo image of the oral cavity of at least one user; and

[0023] - at least one processing means, which receives at least one stereo image; wherein the at least one processing means comprises at least one trained neural network, which analyzes at least one stereo image to estimate at least one depth map; and

[0024] wherein at least one processing means further comprises at least one localization and mapping block, which sequentially integrates at least one stereo image and at least one depth map into the generated three-dimensional model. Petition 870250085553, dated 09 / 22 / 2025, pp. 86 / 108 7 / 19

[0025] The system of the present invention operates according to the passive stereo technique. This means that it uses synchronized image pairs and an algorithm, using trained neural networks, that estimates the depths of the scanned surface, without projecting anything onto it, as is the case with conventional solutions in the state of the art.

[0026] In scenarios with little texture and / or reflections, such as a patient's mouth, traditional passive stereo algorithms generally perform poorly. The poor performance in these scenarios is due to the fact that traditional algorithms search for key points in each image and then attempt to match the key points from one image to another (this is called stereo matching). Because there is little texture and / or reflections, the key points are very ambiguous and the algorithms make many errors in processing.

[0027] The present invention utilizes an algorithm enriched by the use of a highly complex neural network, instead of a traditional stereo matching algorithm. The so-called neural network estimates the depth of images without using explicit keypoints and achieves significantly greater accuracy than traditional stereo matching algorithms. To achieve this accuracy, the neural network must be trained with highly realistic synthetic images and depths.

[0028] In this sense, the algorithm of the present invention, by comprising this neural network, requires a high processing capacity. To avoid making the intraoral device more complex, which would make it more expensive, it is decided to perform said processing outside the device, either in an external electronic device specifically designed for this purpose or on a server located in the cloud. Based on the above, it is possible to solve both problems presented by the application, such as simplifying the hardware used, and also improving the quality of the three-dimensional modeling of the patient's dental prosthesis, which can be obtained in real time by the user, thanks to the work of the neural network.

[0029] According to another embodiment of the invention, at least one processing means further comprises at least one block of Petition 870250085553, dated 09 / 22 / 2025, pp. 87 / 108 8 / 19 post-processing, which eliminates at least one noisy depth from the three-dimensional model and recalculates the pose of at least one set of cameras. This post-processing is performed using all the information captured during the scanning performed by the intraoral device. The recalculation of the pose of at least one set of cameras aims to allow the system of the invention, as images are captured, along with the depth estimation, to estimate the position of at least one set of cameras. This is done to know how to combine the different depths. As the positions of at least one set of cameras are estimated, a cumulative error is also generated, which must be corrected later in the post-processing block.

[0030] According to another embodiment of the invention, the system further comprises at least one receiving device, which receives at least one stereo image from at least one set of cameras, in order to send it to at least one processing medium.

[0031] According to another embodiment of the invention, at least one receiving device is at least one of a computer, a notebook, a tablet, and a smartphone.

[0032] According to another embodiment of the invention, at least one processing means is disposed in at least one receiving device. This allows the user to perform data processing on their own computer, if they have one with the necessary processing capacity.

[0033] According to another embodiment of the invention, at least one receiving device comprises at least one display interface. This display interface allows the user to view images acquired by the intraoral device and / or the resulting three-dimensional model of the dental prosthesis in real time.

[0034] According to another embodiment of the invention, at least one processing medium is disposed of in a cloud. This allows users who do not have the processing capacity necessary to host the processing medium along with the neural network to have an option in Petition 870250085553, dated 09 / 22 / 2025, pages 88 / 108 9 / 19 that the information collected by the intraoral device is sent directly to a cloud server or used as a gateway to the receiving device, which, as mentioned above, can be a computer or a tablet, among others.

[0035] Furthermore, since the processing medium and neural network are located in a cloud, the system can simultaneously analyze information from multiple intraoral devices, sending back the respective 3D models in real time. This feature allows for a further reduction in system cost, as it eliminates the need for a separate processing medium for each operating intraoral device, as is the case with conventional solutions.

[0036] According to another embodiment of the invention, at least one set of cameras comprises at least one first camera and at least one second camera. This allows obtaining the synchronized image pairs necessary for the algorithm and neural network to correctly estimate the depths of the scanned surface.

[0037] According to another embodiment of the invention, at least one intraoral device communicates with at least one wireless processing medium.

[0038] According to another embodiment of the invention, at least one intraoral device communicates with at least one processing medium by means of a communication cable.

[0039] According to another embodiment of the invention, at least one intraoral device communicates with at least one wireless receiving device.

[0040] According to another embodiment of the invention, at least one intraoral device communicates with at least one receiving device by means of a communication cable.

[0041] According to another embodiment of the invention, at least one receiving device communicates with at least one wireless cloud.

[0042] According to another embodiment of the invention, the Petition 870250085553, dated 09 / 22 / 2025, pp. 89 / 108 10 / 19 at least one intraoral device further comprises at least one battery, which allows said intraoral device to operate without the need to be directly connected to a power source, such as a plug.

[0043] On the other hand, according to a second preferred embodiment of the invention, a method for constructing a three-dimensional model of a dental prosthesis for at least one user is also described, comprising the steps of:

[0044] a) capture at least one stereo image through at least one set of cameras arranged in at least one intraoral device;

[0045] b) receive at least one stereo image by at least one processing means;

[0046] c) analyze at least one stereo image to estimate at least one depth map using at least one trained neural network comprising at least one processing means; and

[0047] d) sequentially integrate at least one stereo image and at least one depth map into the three-dimensional model generated by means of at least one localization and mapping block, which comprises at least one processing means.

[0048] As mentioned above, upon capturing the images, the intraoral device begins sending real-time stereo images to a linking device, such as a computer, which then sends this information to a cloud server where the processing medium is located, or sends these images directly to the processing medium located on the computer. Once the images are available for analysis by the processing medium, the neural network trained specifically for this task analyzes each stereo image sent and estimates a depth map from them.

[0049] The depth map, along with the stereo image, are integrated into the reconstruction generated up to that point by means of at least one location and mapping block, which compares the information. Petition 870250085553, dated 09 / 22 / 2025, pp. 90 / 108 11 / 19 received with the partial reconstruction of the scene and provides a pose for the stereo image.

[0050] According to another embodiment of the invention, the method further comprises removing at least one noisy depth from the three-dimensional model and recalculating the pose of at least one set of cameras through at least one post-processing block, which comprises at least one processing means.

[0051] According to another embodiment of the invention, the method further comprises, before step c), receiving at least one stereo image, from at least one set of cameras, by at least one receiving device, and then sending it to at least one processing medium.

[0052] According to another embodiment of the invention, the method further comprises displaying the information sent and received from at least one processing medium by means of at least one display interface disposed in at least one receiving device.

[0053] According to another embodiment of the invention, the method further comprises generating the three-dimensional model of the dental prosthesis of at least one user in real time and sending it to at least one display interface.

[0054] Finally, according to a third preferred embodiment of the invention, a computer-readable storage medium is also described, comprising instructions that, when executed by at least one processor, cause the at least one processor to execute the method for constructing a three-dimensional model of a dental prosthesis for at least one user.

[0055] From the foregoing, it is possible to observe that an important difference between the present invention and prior art solutions is related to the fact that the present system, instead of resorting to techniques such as active stereomicroscopy, structured light or confocal microscopy, uses a powerful neuronal model that allows it to estimate depths. Petition 870250085553, dated 09 / 22 / 2025, pp. 91-108 12 / 19 using exclusively the passive stereomicroscope technique. This means that the intraoral device sensors are considerably less complex and expensive, as they consist of at least one common camera and a circuit to synchronize them.

[0056] Furthermore, both the calculation of stereo image depths and the reconstruction and post-processing process can be performed on a server located in the cloud, where the processing medium is located along with the trained neural network. In this case, the dentist's computer only sends the information obtained by the intraoral device to this data cloud and allows the reconstruction status to be viewed in real time on a display interface, allowing the dentist to control the process.In contrast, traditional scanners perform reconstruction and post-processing on the dentist's own computer, which requires the computer to have sufficiently powerful hardware to host the processing algorithms used. Often, the algorithm used by these solutions must be adapted to this type of computer, directly affecting the quality of the resulting three-dimensional model. This does not occur in the present invention, where the quality of the model is improved because there is no such limitation in terms of processing capacity.

[0057] Finally, none of the state-of-the-art solutions offer the ability to use a single processing medium to analyze data sent by multiple intraoral devices, which optimizes resource utilization and reduces the costs associated with system implementation, allowing greater accessibility to this type of technology for dentists and patients. BRIEF DESCRIPTION OF THE FIGURES

[0058] As part of the present invention, the following representative figure is presented, which shows a preferred configuration of the invention and, therefore, should not be considered as limiting the definition of the claimed matter.

[0059] Figure 1 shows a block diagram of Petition 870250085553, dated 09 / 22 / 2025, pp. 92 / 108 13 / 19 Intraoral scanning process, according to the state of the art.

[0060] Figure 2 shows a general scheme of the passive stereo technique, according to a preferred configuration of the invention.

[0061] Figure 3 shows a block diagram of the intraoral scanning process, according to a preferred configuration of the present invention. DETAILED DESCRIPTION OF THE FIGURES

[0062] With reference to the attached figures, Figure 1 shows a block diagram of the traditional intraoral scanning process to obtain a three-dimensional representation of a patient's dental prosthesis. Specifically, a first data capture and reconstruction stage (1) is observed, which is composed of two sub-stages. The first sub-stage (1a) aims at the actual scanning performed by a skilled professional, such as a dentist, in the patient's mouth. This is done using a physical device that is inserted into the user's mouth, which includes at least one data capture device.During scanning, the scanner sends the captured data in real time to the dentist's computer, where these scanners generally operate under the theoretical principles of confocal or structured light microscopy (active stereo), which are usually images, depths, and measurements from an Inertial Measurement Unit (IMU) comprising an accelerometer and a gyroscope to measure angular velocities and accelerations. Once the information obtained by the scanner is sent to the technician's computer, it sequentially integrates the images, depths, and other measurements to estimate the camera pose in each image and thus construct the 3D model (substep (1b)). In this way, a first 3D model of the patient's dental prosthesis is obtained, still imprecise and dirty (block (2)).

[0063] As the 3D model obtained in this first stage is not ideal for use in dental treatments, a post-processing stage (3) is necessary, where this first 3D model is cleaned by eliminating noisy points or points that do not correspond to the dental prosthesis (sub-stage (3a)). After that, Petition 870250085553, dated 09 / 22 / 2025, pp. 93 / 108 14 / 19 The dentist's computer recalculates the reconstruction using all the information received during the scan (substep (3b)). Finally, the poses and depths are optimized to minimize the reprojection error (substep (3c)), which corresponds to the difference between the captured images and the images generated from the reconstruction of the 3D model, after which a corrected and clean 3D model is obtained (block (4)), which allows its use by the skilled professional to develop a specific treatment for the patient (block (5)), among which we can mention invisible aligners, relaxation planes, dental crowns, among others.

[0064] It should be noted that these state-of-the-art techniques require that the user have a computer (10) with high processing capabilities, given the number of images to be processed, and that it is able to execute the processing algorithms necessary to generate the 3D model, in which at least the steps described by blocks (1), (2) and (3) must be performed.

[0065] On the other hand, Figure 2 presents a general scheme of the passive stereo technique used by the present invention, in which it is possible to understand the operation of the intraoral scanner or device (30) of the technology. In this sense, in the embodiment presented in Figure 2, the intraoral device (30) comprises two left and right cameras (12a, 12b), positioned on a baseline (11), wherein said cameras (12a, 12b) in turn comprise a left and right lens (13a, 13b), respectively. The left and right cameras (12a, 12b) can be positioned parallel to the baseline (11) or at an angle to it.

[0066] Each camera (12a, 12b) of the intraoral device (30) forms an image plane (14a, 14b), through which the real points (15, 16) in each image obtained by the cameras (12a, 12b) are visualized. As mentioned earlier, the intraoral device (30) uses the passive stereo technique to estimate the depth (17) of the scanned surface, which is obtained through synchronized image pairs and an algorithm that processes the images, without the need to project anything onto them, as occurs in other state solutions. Petition 870250085553, dated 09 / 22 / 2025, pp. 94 / 108 15 / 19 of the technique to solve this problem.

[0067] This is especially relevant in scenarios with little texture and / or reflections, such as a patient's mouth, where traditional passive stereo algorithms generally perform poorly because these traditional algorithms search for key points in each image and then attempt to relate these key points from one image to those of another (stereo matching). Because the inside of the mouth has surfaces that are mostly poorly textured and / or reflective, the key points are very ambiguous and difficult to locate, so these traditional algorithms frequently make mistakes, providing inaccurate results.

[0068] The present invention, by using a neural network instead of a traditional algorithm for stereo matching, does not require explicit location of key points to calculate the depth of the images, which improves accuracy in situations with little texture and / or reflections, such as the inside of a patient's mouth. To achieve this accuracy, the neural network must be trained with highly realistic synthetic images and depths.

[0069] Finally, Figure 3 shows a block diagram of the intraoral scanning process, according to a preferred embodiment of the invention. Specifically, a first data capture step (100) is shown, which is divided into two substeps. The first aims at capturing synchronized image pairs (stereo images) by the person skilled in the art, through the intraoral scanner or device, and subsequent transmission to their computer, via wireless or data cable (substep (100a)). The computer then transmits the stereo images in real time to a server or cloud (substep (100b)). The intraoral device preferably consists of two cameras and a circuit that synchronizes the capture of both cameras, where the depths are calculated by a neural network from these images, which preferably resides in the same cloud where the images are received from the intraoral device.As a result of this first data capture step (100), a stream of stereo images (block (200)) is obtained from the intraoral device to the computer of the person skilled in the art and then to the. Petition 870250085553, dated 09 / 22 / 2025, pages 95 / 108 16 / 19 cloud.

[0070] This is an important difference between what is described by the prior art and the present invention, since, as can be observed, the invention system only requires the use of the skilled person's computer (10) in this data capture step (100), after which the information is processed in the cloud, where the processing algorithms are hosted, together with the neural network. This avoids the skilled person needing to have a computer with high processing power, which could be achieved even with the use of an electronic device capable of connecting to the cloud, such as a tablet or smartphone.

[0071] After this first stage of data capture and sending of information to the cloud (100), in which the stereo image stream is obtained (200), a cloud reconstruction stage follows (300), which comprises a sub-stage of depth map calculation (300a) for each pair of images, using a neural network, and a sequential integration sub-stage for 3D reconstruction of the depth maps and images (300b). The neural network compares the two images (RGB, for example) and estimates the depth of each pixel from the relative movement of objects between the two images. In contrast, traditional scanners estimate depth using a confocal microscope or structured light, which are less precise, requiring surfaces with appreciable texture. As a result of this cloud reconstruction stage (300), an imprecise and messy 3D model is obtained (block (400)).

[0072] Once the first 3D model (400) is obtained, a post-processing step must be performed in the cloud (500), comprising three sub-steps. The first aims at cleaning the 3D reconstruction (500a) by removing noisy points or points that do not correspond to the dental prosthesis, after which the processing medium recalculates the reconstruction (500b), using all the information received during the scanning and data capture step (100). Finally, the poses and depths (500c) are optimized to minimize the reprojection error, which, as mentioned above, corresponds to the difference between the captured images. Petition 870250085553, dated 09 / 22 / 2025, pages 96 / 108 17 / 19 and the images generated from the reconstruction.

[0073] As a result of these two stages of reconstruction and post-processing in the cloud, a corrected and clean 3D model is obtained (block (600)), with improved accuracy compared to the state of the art, where the skilled in the art can download the generated 3D model from the cloud to perform dental treatment (700), such as invisible aligners, relaxation planes, dental crowns, among others.

[0074] In this sense, it is important to highlight the relevance of the fact that the reconstruction (300) and post-processing (500) steps are performed in a cloud (20), where the processing means are hosted, which comprise the trained neural network and the algorithms that make up the localization and mapping block, which sequentially integrates at least one stereo image and at least one depth map into the generated three-dimensional model. This not only avoids the skill holder having to depend on a device with high processing power, as mentioned above, but also allows cloud-based processing tools with greater capacity than those used in the state of the art, which are limited by the processing capacity of the intraoral device or the skill holder's computer.This results in a solution where the expert receives the results with the corrected and cleaned 3D model in real time, thanks to the high processing power and neural network that allows for more accurate results in less time.

[0075] Finally, it is also worth highlighting the possibility of the inventive system operating with various intraoral devices through a single cloud, which allows for a significant reduction in the costs associated with the implementation of the system, making it accessible to most skilled individuals, for whom only one processing means is necessary, which is not the case with prior art solutions, for which it is necessary to pay for the complete processing system each time a skilled individual acquires the product. NUMERICAL REFERENCES Petition 870250085553, dated 09 / 22 / 2025, pages 97 / 108 18 / 19

[0076]

[0077] data for the computer

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099] data to the computer

[0100] real to the cloud

[0101]

[0102] Data capture and reconstruction: Digitization by a specialist skilled in the technique and sending of the data; 3D model reconstruction. Inaccurate and messy 3D model Post-processing 3a Cleaning the inaccurate and dirty 3D model 3b Recalculate the reconstruction 3c Optimization of poses and depths Corrected and cleaned 3D model Treatment Computer expert Baseline 12th Chamber left 12b Right camera 13th Left Lens 13b Right lens 14a Left image plane 14b Right image plane, 16 Real point Depth Cloud Intraoral device 100 Data Capture 100th Digitization by someone skilled in the technique and sending 100b Stereo image transmission in real time 200 Stereo image stream 300 Cloud Reconstruction Petition 870250085553, dated 09 / 22 / 2025, pages 98 / 108 19 / 19

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111] 300a Depth map calculation 300b Sequential integration for 3D reconstruction 400 Inaccurate and dirty 3D model 500 Cloud post-processing 500a Cleaning of inaccurate and dirty 3D model 500b Recalculate reconstruction 500c Optimization of poses and depths 600 Corrected and cleaned 3D model 700 Treatment Petition 870250085553, dated 09 / 22 / 2025, pp. 99 / 108

Claims

1 / 4 CLAIMS 1. A system for constructing a three-dimensional model of a dental prosthesis for at least one user, characterized in that it comprises: - at least one intraoral device, to be placed in the oral cavity of at least one user; - at least one set of cameras arranged in at least one intraoral device, to capture at least one stereo image of the oral cavity of at least one user; and - at least one processing means, which receives at least one stereo image; wherein the at least one processing means comprises at least one trained neural network, which analyzes at least one stereo image to estimate at least one depth map; and wherein the at least one processing means further comprises at least one localization and mapping block, which sequentially integrates at least one stereo image and at least one depth map into the generated three-dimensional model.

2. System, according to claim 1, characterized in that at least one processing means further comprises at least one post-processing block, which eliminates at least one noisy depth from the three-dimensional model and recalculates the pose of at least one set of cameras.

3. A system, according to any one of claims 1 to 2, characterized in that it further comprises at least one receiving device, which receives at least one stereo image from at least one set of cameras, in order to send it to at least one processing medium.

4. System, according to claim 3, characterized in that at least one receiving device is at least one of a computer, a notebook, a tablet, and a smartphone. Petition 870250085553, dated 09 / 22 / 2025, pp. 100 / 108 2 / 4 5. A system, according to any one of claims 3 and 4, characterized in that at least one processing means is disposed in at least one receiving device.

6. A system according to any one of claims 3 to 5, characterized in that at least one receiving device comprises at least one display interface.

7. A system, according to either of claims 1 and 2, characterized in that at least one processing medium is located in a cloud.

8. System, according to any one of claims 1 to 7, characterized in that at least one set of cameras comprises at least one first camera and at least one second camera.

9. System, according to any one of claims 1 to 8, characterized in that at least one intraoral device communicates with at least one wireless processing medium.

10. System, according to any one of claims 1 to 9, characterized in that at least one intraoral device communicates with at least one processing medium by means of a communication cable.

11. System, according to any one of claims 3 to 10, characterized in that at least one intraoral device communicates with at least one wireless receiving device.

12. System, according to any one of claims 3 to 10, characterized in that at least one intraoral device communicates with at least one receiving device by means of a communication cable.

13. System, according to any one of claims 7 to 12, characterized in that at least one receiving device communicates with at least one wireless cloud.

14. System, according to any one of claims 1 to 13, characterized in that at least one intraoral device further comprises at least one battery. Petition 870250085553, dated 22 / 09 / 2025, pp. 101 / 108 3 / 4 15. Method for constructing a three-dimensional model of a dental prosthesis for at least one user, according to the system as defined in claims 1 to 14, characterized in that it comprises the steps of: a) capturing at least one stereo image through at least one set of cameras arranged in at least one intraoral device; b) receiving at least one stereo image by at least one processing means; c) analyzing at least one stereo image to estimate at least one depth map through at least one trained neural network, comprising at least one processing means; d) sequentially integrating at least one stereo image and at least one depth map into the generated three-dimensional model by means of at least one localization and mapping block, comprising at least one processing means.

16. Method, according to claim 15, characterized in that it further comprises eliminating at least one noisy depth from the three-dimensional model and recalculating the pose of at least one set of cameras through at least one post-processing block, which comprises at least one processing means.

17. A method according to any one of claims 15 and 16, characterized in that it further comprises, before step c), receiving at least one stereo image from at least one set of cameras by at least one receiving device, and subsequently sending it to at least one processing medium.

18. Method, according to claim 17, characterized in that it further comprises displaying the information sent and received from at least one processing medium through at least one display interface disposed in at least one receiving device.

19. Method, according to any of the claims 15 Petition 870250085553, dated 09 / 22 / 2025, pp. 102 / 108 4 / 4 to 18, characterized in that it further comprises generating the three-dimensional model of the dental prosthesis of at least one user in real time and sending it to at least one display interface.

20. Computer-readable storage medium, characterized in that it comprises instructions which, when executed by at least one processor, cause that processor to execute the method as defined in any one of claims 15 to 19. Petition 870250085553, dated 09 / 22 / 2025, pp. 103 / 108

Images