3D Printing Feature Classification via ML Intermediary
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D printing technologies lack the ability to automatically detect and classify critical features in 3D part geometry, which are essential for ensuring structural integrity and accuracy, and do not effectively utilize user input for training machine learning models to enhance feature recognition.
Innovation Solution
A system comprising a processor and memory that uses machine learning models to identify and classify features in 3D part geometry, incorporating user feedback to improve detection and classification accuracy, and allows for the generation of production data based on these classifications, while also accommodating user-specific data privacy by isolating user data for training models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If machine learning models are used to automatically detect and classify critical features in 3D part geometry, then manufacturing precision and reliability are improved, but device complexity increases
Solution Approach 1:
A feature classification service acts as an intermediary between the 3D printing system and machine learning models. This service receives 3D part geometry data, orchestrates multiple machine learning models (detectron2, pointnet, resnet), and returns classified critical features. The intermediary abstracts the complexity of running multiple ML models, allowing the 3D printing system to benefit from improved feature detection accuracy without directly managing the complex ML infrastructure.
2Measurement precision
If user feedback is collected and used to train machine learning models, then measurement precision and manufacturing precision are improved, but loss of information and privacy concerns increase
Solution Approach 1:
The system extracts only the necessary feature classification feedback from user interactions while preserving user data privacy. Instead of collecting and storing complete user datasets, the system extracts specific classification corrections and feedback signals needed to improve model accuracy. This selective extraction reduces information loss and privacy risks while maintaining measurement precision improvements.
Solution Approach 2:
The system implements a feedback mechanism where user corrections of feature classifications are captured and used to retrain machine learning models. The feedback service receives user feedback on detected features, processes this information to update training datasets, and retrains models to improve future classification accuracy. This closed-loop feedback system continuously improves measurement precision while managing user data through controlled feedback collection.
3Manufacturing precision
If multiple machine learning models are executed to classify features, then manufacturing precision is improved, but productivity decreases due to increased processing time
Solution Approach 1:
The feature classification process is segmented into multiple specialized machine learning models, each targeting specific types of features or classification tasks. Models such as detectron2, pointnet, and resnet are divided and assigned to different feature detection tasks. This segmentation allows parallel execution of models and enables the system to process different feature types simultaneously, improving overall processing efficiency while maintaining high classification accuracy through specialized model expertise.
Solution Approach 2:
The system performs preliminary feature detection and classification using multiple machine learning models before the actual 3D printing production. By executing all necessary classification operations in advance and caching results, the system avoids reprocessing during production cycles. The feature classification service pre-processes 3D part geometry data, identifies critical features, and stores classification results for rapid retrieval during manufacturing, thereby maintaining high precision without compromising production productivity.
Data Source
AI summary
A 3D printing apparatus and method determines features based on design data of an object and at least one classification of a determined feature. The at least one classification includes a classification that the determined feature is a critical feature for the object. At least one print setting for forming the determined feature is modified based on the classification that the determined feature is a critical feature.