VR Labeling of 3D Models for ML Training Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for generating labeled training data for machine learning and artificial intelligence face challenges such as time intensity, user error, and difficulty in navigating 3D environments, particularly in accurately labeling and verifying objects in 3D models.
Innovation Solution
A system and method utilizing virtual reality (VR) and augmented reality (AR) to enable users to interactively label 3D models by displaying a VR environment, allowing users to select and label objects through user interaction, and generating a labeled 3D model by embedding the labeling data into the model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual labeling of 3D models is performed using conventional systems, then labeling accuracy can be maintained, but the process becomes time intensive and inefficient
Solution Approach 1:
The patent uses virtual reality to create a digital copy of the 3D environment that users can navigate and label more efficiently. The VR system replicates the physical 3D space, allowing users to interact with and label objects in a simulated environment, which reduces the time required compared to traditional manual labeling methods while maintaining accuracy through immersive visualization
Solution Approach 2:
The patent replaces conventional mechanical interaction methods (such as viewing 3D models on 2D screens or using traditional navigation tools) with virtual reality technology. This substitution enables more intuitive and efficient labeling through immersive 3D interaction, spatial awareness, and natural user interfaces, thereby reducing labeling time without sacrificing precision
2Productivity
If users manually label objects in 3D models using conventional interfaces, then labeling can be performed, but user error increases due to difficulty in navigating 3D environments
Solution Approach 1:
The VR system creates an immersive digital replica of the 3D environment, allowing users to navigate and label objects in a familiar spatial context. This copying approach maintains the spatial relationships and perspectives of the original 3D model while providing enhanced interaction capabilities that reduce user error and improve labeling reliability
Solution Approach 2:
The patent introduces virtual reality as an intermediary between the user and the 3D model. This intermediary provides intuitive navigation tools, spatial anchoring, and enhanced visualization that help users accurately identify and label objects in complex 3D environments, thereby reducing errors while maintaining productivity
3Ease of manufacture
If conventional systems are used for labeling, then the process can be completed with basic equipment, but the systems become complex and difficult to operate
Solution Approach 1:
The patent replaces complex conventional labeling interfaces with virtual reality technology, which simplifies user interaction through immersive 3D navigation and intuitive controls. The VR system handles the complexity of 3D environment representation and object identification automatically, presenting users with simple, natural interaction methods while maintaining ease of operation
Data Source
AI summary
A computer-implemented method for labeling a three-dimensional (3D) model using virtual reality (VR) techniques implemented by a computer system including a processor is provided herein. The method may include (i) receiving a 3D model including an environmental feature that is unlabeled, (ii) displaying, through a VR device in communication with the processor, a VR environment to a user representing the 3D model, (iii) prompting a user to input labeling data for the environmental feature displayed within the VR environment of the VR device by prompting the user to select the environmental feature through user interaction with the VR device, and input labeling data for the environmental feature, wherein the labeling data identifies the environmental feature, and (iv) generating a labeled 3D model by embedding the labeling data associated with the selected environmental feature into the 3D model.