Fluoroscopic Object Pose Estimation via Inertia Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object identification and object semantic segmentation systems require a large number of projection images to generate a complete sinogram, which is time-consuming and resource-intensive.
Innovation Solution
The system uses a minimal number of fluoroscopic images to simulate a complete sinogram by determining moments of inertia and principal axes, allowing for the reconstruction of objects and estimation of their initial poses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large number of projection images are used to generate a complete sinogram, then the completeness and accuracy of object reconstruction is improved, but the time consumption and resource requirements increase
Solution Approach 1:
The system performs preliminary detection of object geometric properties (moments of inertia, principal axes) from a minimal set of fluoroscopic images before generating the complete sinogram. This preliminary action allows the system to simulate additional projection angles based on detected geometric characteristics, reducing the need for acquiring numerous actual projection images while maintaining reconstruction accuracy.
Solution Approach 2:
The system changes the parameter set used for sinogram generation by incorporating moments of inertia and principal axes as key geometric parameters. Instead of relying solely on multiple projection images, the system uses these geometric parameters to simulate missing projection angles, thereby achieving complete sinogram generation with fewer actual images and reduced processing time.
2Reliability
If a large number of projection images are captured to ensure complete sinogram generation, then the quality of sinogram data is improved, but the resource intensity and complexity of the system increase
Solution Approach 1:
The system creates simulated copies of projection images at unselected angles by using detected geometric properties (moments of inertia, principal axes) from a minimal set of actual fluoroscopic images. These simulated images serve as copies that complete the sinogram without requiring acquisition of numerous additional physical images, thereby maintaining data quality while reducing system resource requirements.
Solution Approach 2:
The system achieves multi-functionality by using the same detected geometric properties (moments of inertia, principal axes) for multiple purposes: characterizing the object, simulating additional projection angles, and completing the sinogram. This universal use of geometric parameters eliminates the need for separate acquisition processes for each projection angle, reducing overall system complexity and resource requirements.
3Measurement precision
If multiple fluoroscopic images are used to determine geometric properties, then the accuracy of object characterization is improved, but the number of images required increases processing time
Solution Approach 1:
The system applies partial action by using only a minimal necessary set of fluoroscopic images to detect geometric properties (moments of inertia, principal axes) rather than processing all possible projection images. This partial processing approach maintains sufficient geometric characterization accuracy while significantly improving processing efficiency compared to analyzing complete image sets.
Data Source
AI summary
A system includes an imaging device configured to rotate axially about a rotational axis and to capture a plurality of images associated with a sinogram at a plurality of angles, respectively, along the rotational axis, one or more processors, and a memory storing instructions that, when executed by the one or more processors, cause the system to: detect an object in each image of the plurality of images, determine a moment of inertia associated with the object based on the plurality of images, simulate one or more images for the sinogram based at least in part on the moment of inertia, generate a reconstruction of the object based on the sinogram; and determine one or more properties of the object based on the reconstruction.