3D Object Segmentation and Editing From Multi-View 2D Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional 3D object segmentation techniques face challenges in precise segmentation of complex shapes, require manual user intervention, and are computationally intensive, making real-time processing and user-friendly interaction difficult.
Innovation Solution
A method for 3D object segmentation that involves obtaining 2D images from multiple viewpoints, generating 3D images with segmented objects, and allowing user interaction through a UI for selecting and editing objects in a hierarchical structure, enabling precise customization and intuitive editing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional 3D object segmentation techniques are used, then objects can be segmented in a 3D scene, but the detailed structure of the object cannot be precisely segmented
Solution Approach 1:
The patent applies segmentation by dividing the 3D object into multiple 2D images from different viewpoints. Each 2D image is then segmented independently using a segmentation model, and the results are integrated to form the complete 3D object segmentation. This approach enables precise segmentation of detailed structures while maintaining system feasibility.
2Measurement precision
If manual segmentation threshold setting is required, then segmentation can be performed, but user workload increases and accuracy depends on user expertise
Solution Approach 1:
The patent implements self-service by enabling the segmentation model to automatically determine optimal segmentation thresholds without requiring manual user input. The system processes the 2D images and performs segmentation autonomously, eliminating the need for users to manually set parameters or have specialized expertise in segmentation techniques.
3Measurement precision
If complex preprocessing or postprocessing steps are adopted, then object segmentation can be achieved, but real-time processing becomes difficult
Solution Approach 1:
The patent extracts and processes only the essential features from the 2D images that are necessary for accurate segmentation, avoiding unnecessary complex preprocessing or postprocessing steps. By focusing on key features and using an efficient segmentation model, the system achieves both high segmentation quality and real-time processing capability.
4Measurement precision
If a network structure with large computational volume is used, then segmentation accuracy can be improved, but real-time processing becomes difficult
Solution Approach 1:
The patent applies partial action by processing a limited set of key features from the 2D images rather than analyzing all possible image data. This selective processing approach maintains segmentation accuracy while significantly reducing computational energy consumption, enabling real-time performance.
Data Source
AI summary
A device according to one embodiment of the present disclosure may comprise a display; and at least one processor configured to: obtain a plurality of two-dimensional (2-Dimensional) images corresponding to a plurality of viewpoints, obtain a 3D image including one or more segmented 3D objects based on the obtained plurality of 2D images, display a UI (User Interface) element indicating that a segmented 3D object is selected on the display in response to a first command to select one of the one or more segmented 3D objects representing a unit object or a detailed object constituting the unit object, and display an edited 3D object on the display in response to a second command to edit the selected 3D object.