Search

JP-7856450-B2 - Machine Learning for 3D Segmentation

JP7856450B2JP 7856450 B2JP7856450 B2JP 7856450B2JP-7856450-B2

Inventors

  • アリアン ジョーダン
  • エロイ メア

Assignees

  • ダッソー システムズ

Dates

Publication Date
20260511
Application Date
20220303
Priority Date
20210310

Claims (12)

  1. A computer implementation method for machine learning, - A step of providing a dataset of training samples, wherein each training sample includes a pair of parts of a 3D modeled object, and each pair of parts is labeled with a value indicating whether the two parts belong to the same segment of the 3D modeled object, and - A step of training a neural network based on the dataset, wherein the neural network is configured to receive two parts of a 3D modeled object representing a machine part as input and to output individual values indicating the extent to which the two parts belong to the same segment of the 3D modeled object, thereby making the neural network available for 3D segmentation. Includes , The step of training the neural network includes minimizing a loss that includes a term that penalizes the error between the respective values indicating whether two parts of a given pair belong to the same segment of a 3D modeled object and the individual values that the neural network outputs for the two parts, for each pair of the dataset. The neural network converts the two input parts into corresponding intermediate features, and the loss further includes another term that penalizes the error between the respective values indicating whether the two parts of a given pair belong to the same segment of a 3D modeled object and the approximation between the intermediate features corresponding to the two parts, for each pair of the dataset. method.
  2. The aforementioned neural network, - An encoding neural network configured to receive a portion of a 3D modeled object as input and encode it into a latent vector, and - Includes a predictive neural network configured to receive a pair of latent vectors output by the encoding neural network as input, and to output individual values indicating the extent to which the two parts encoded by the latent vectors belong to the same segment of the 3D modeled object , The method according to claim 1.
  3. The encoding neural network is configured to receive a portion of a 3D modeled object and an adjacent portion of its input as input . The method according to claim 2.
  4. The above clause is as follows: Here, (( X1 (1) , X2 (2) ), ..., ( XN (1) , XN (2) )) are pairs of the dataset, and g w (f w ( Xi (1) ), f w ( Xi (2) )) are the individual values output by the neural network for the pair ( Xi (1) , Xi (2) ), is the neural network, pi is the respective value indicating whether the parts Xi (1) and Xi (2) belong to the same segment of the 3D modeled object, and dCE is a loss that penalizes the error between gw ( fw ( Xi (1) ), fw ( Xi (2) )) and pi . The aforementioned other item is as follows: Here, f w (X i (1) ) and f w (X i (2) ) are the intermediate features corresponding to the parts Xi (1) and Xi (2 ) of the pair (X i (1) , Xi (2)) , and d M is a loss that penalizes the error between p i and the approximation between f w (X i (1) ) and f w (X i (2) ) . The method according to claim 1 .
  5. The step of providing the aforementioned dataset is, - A step of providing an initial dataset of segmented 3D model objects, and - For each segmented 3D model object in the initial dataset, a step of creating a pair of parts of the segmented 3D model object, wherein, of the pairs, the parts of a pair belong to either the same segment or different segments of the segmented 3D model object, and each pair includes a value indicating whether the two parts belong to the same segment of the 3D model object . The method according to any one of claims 1 to 4, including the method described in any one of claims 1 to 4 .
  6. The step of creating the pair includes creating a pair of parts with respect to one or more segmented 3D modeled objects in the initial dataset, where at least one of the parts does not belong entirely to a single segment . The method according to claim 5 .
  7. The step of providing the dataset further includes the step of creating pairs of minibatches, and the training step is performed for each of the minibatches . The method according to claim 5 or 6 .
  8. A computer implementation method for using a trainable neural network according to any one of claims 1 to 7 for 3D segmentation, - A step of providing a 3D model object representing a machine part, - The step of applying the neural network to pairs of parts of the 3D modeled object, thereby outputting a set of values for each pair of parts of the 3D modeled object, wherein each value indicates the extent to which the two parts belong to the same segment of the 3D modeled object, - A step of performing segmentation of the 3D modeled object based on a set of one or more similarity indices between adjacent parts of the 3D modeled object, wherein the set of similarity indices includes the set of values , Methods that include...
  9. The step of performing the aforementioned segmentation is, - A step of performing segmentation based on the network of the 3D modeled object based on the set of values, and then performing another segmentation of the 3D modeled object based on at least one other similarity metric, or a step of performing the other segmentation, and then performing segmentation based on the network, or - A step of performing segmentation of the 3D model object based solely on the set of the aforementioned values, or - The step of performing segmentation of a 3D modeled object based on the set of values combined with at least one other similarity metric , The method according to claim 8 .
  10. A computer program comprising instructions for performing the method according to any one of claims 1 to 7 and/or the method according to any one of claims 8 to 9 .
  11. A device comprising a computer - readable data storage medium on which the computer program of claim 10 is recorded.
  12. The system further comprises a processor coupled with the aforementioned computer-readable data storage medium . The device according to claim 11 .

Description

This disclosure relates to the field of computer programs and systems, more specifically to machine learning methods, systems, and programs for 3D segmentation. The market offers numerous systems and programs for the design, engineering, and manufacturing of objects. CAD stands for Computer-Aided Design, and refers to software solutions for designing objects, for example. CAE stands for Computer-Aided Engineering, and refers to software solutions for simulating the physical behavior of future products, for example. CAM stands for Computer-Aided Manufacturing, and refers to software solutions for defining manufacturing processes and operations, for example. In such computer-aided design systems, graphical user interfaces play a crucial role in terms of technical efficiency. These technologies can be integrated into Product Lifecycle Management (PLM) systems. PLM is a business strategy that helps companies share product data, apply common processing, and leverage enterprise knowledge to help develop products from concept to lifecycle, across the concept of an extended enterprise. Dassault Systèmes' PLM solutions (under the trademarks CATIA, ENOVIA, and DELMIA) provide an engineering hub for organizing product engineering knowledge, a manufacturing hub for managing manufacturing engineering knowledge, and an enterprise hub that enables enterprise integration and connectivity to the engineering and manufacturing hubs. Combined, this system provides an open object model that links products, processes, and resources to enable dynamic, knowledge-based product creation and decision support, facilitating product definition, manufacturing preparation, production, and service optimization. This method is illustrated as an example.This method is illustrated as an example.This method is illustrated as an example.This method is illustrated as an example.This method is illustrated as an example.This method is illustrated as an example.This method is illustrated as an example.An example of the system is shown. A computer implementation method for machine learning is provided. This method includes the step of providing a dataset of training samples. Each training sample contains pairs of 3D modeled object parts labeled with their respective values (i.e., each pair is labeled with its respective value). Each value indicates whether the two parts belong to the same segment of the 3D modeled object. The method further includes the step of training a neural network based on this dataset. The neural network is configured to take two parts of a 3D modeled object representing a machine part as input and output distinct values. These distinct values indicate the degree to which the two parts belong to the same segment of the 3D modeled object. This allows the neural network to be used for 3D segmentation. This method is sometimes referred to as a "training method." This learning method constitutes an improved solution for 3D segmentation. In particular, this learning method trains a neural network that can be used for 3D segmentation. In other words, the neural network constitutes a tool for 3D segmentation, and can therefore be used in 3D segmentation methods. A neural network is configured and learns to take any pair of parts of a 3D modeled object representing a machine part as input and to output a distinct value indicating the degree to which the two parts belong to the same segment of the 3D modeled object. In other words, the output of the neural network indicates the degree to which the two parts can be classified as belonging to the same segment of the 3D modeled object, for example, a probability or confidence score that the two parts belong to the same segment. The term "segment" is understood as a segmentation segment, i.e., a geometrically consistent surface portion of the 3D modeled object. Therefore, the output of the neural network can be directly used in segmentation algorithms or methods that use information about the degree to which two adjacent parts of a given pair belong to the same segment, for (e.g., all) pairs of adjacent parts of the 3D modeled object being segmented. Specifically, many segmentation algorithms or methods perform segmentation of a 3D modeled object based on a set of one or more similarity indices between adjacent parts of the 3D modeled object, where each similarity indice includes a set of values indicating how much two adjacent parts of the 3D modeled object belong to the same segment, i.e., a set of shape similarity values between adjacent parts of the 3D modeled object. A neural network may be applied to (e.g., all) pairs of adjacent parts of the 3D modeled object, and this application causes the neural network to output distinct values for each pair, each indicating how much the two parts of that pair belong to the same segment of the 3D modeled object. These distinct values together form a set of values that constitutes a similarity indice, and segmentation algorithms or methods may b