Imaging System Using SLAM Surface Model for Probe Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current imaging techniques, such as ultrasound imaging, face limitations in field of view, accuracy, and cost due to reliance on external hardware and registration methods, which result in errors, slow processing, and insufficient submillimeter accuracy for applications like panoramic ultrasound and SPECT imaging, and lack positional relationship to anatomical features for interventions.
Innovation Solution
An imaging system that uses a camera to acquire surface images and texture simultaneously, employing a SLAM algorithm to generate a surface model for accurate probe localization and image stitching, eliminating the need for additional hardware and enabling precise alignment and tracking of probe position and orientation, allowing for the creation of panoramic images with minimal data transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If imaging techniques are used to stitch images together, then panoramic images can be formed, but the registered images are prone to errors, have large distortions, and are slow to produce
Solution Approach 1:
The patent introduces a surface model as an intermediary element that mediates between multiple images. The surface model is constructed from images of the subject's surface and serves as a reference framework for stitching. By using this intermediary surface model, the system achieves accurate alignment of multiple images without the errors and distortions associated with traditional registration methods
Solution Approach 2:
The patent replaces traditional mechanical registration methods (such as speckle correlation algorithms) with a geometric approach based on surface modeling. Instead of relying on complex image processing algorithms to align images, the system uses the geometric relationships captured in the surface model to determine probe positions and orientations, thereby eliminating registration errors and distortions
2Measurement precision
If external tracking devices with pre-calibrated cameras are used, then image registration accuracy improves, but the system becomes very costly and requires additional hardware
Solution Approach 1:
The system uses the imaging probe itself to capture surface images, which are then used to construct the surface model for tracking. The same probe that images the internal structure also images the surface, eliminating the need for separate tracking cameras or external hardware. This self-service approach provides accurate position measurement without additional costly equipment
Solution Approach 2:
The imaging probe is given multiple functions: it both images the internal structure of interest and captures surface images for localization. By making the probe universal and multi-functional, the system eliminates the need for separate tracking devices while maintaining accurate position measurement capabilities
3Speed
If optical measurement devices are used for tracking, then measurement speed improves, but accuracy in measuring rotations is limited and insufficient for submillimeter accuracy
Solution Approach 1:
The patent transitions from 2D optical measurement to 3D surface modeling. By constructing a three-dimensional surface model from multiple surface images, the system captures both positional and rotational information in three dimensions. This dimensional upgrade provides the submillimeter accuracy required for rotation measurement while maintaining high measurement speed
4Area of stationary object
If panoramic images are produced without known positional relationship to anatomical features, then large field of view is achieved, but it becomes difficult to locate correct entry points for interventions
Solution Approach 1:
The patent merges the internal structure images with the surface model into a unified coordinate system. The surface model provides anatomical landmarks and positional references that are directly tied to the internal structures. This merging allows operators to easily correlate surface features with internal structures, making it simple to locate correct entry points for interventions while maintaining the large field of view
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
An imaging system combines a probe for obtaining image data in respect of a structure of interest below a surface of a subject, a first camera for obtaining images of the surface of the subject and a second camera for capturing images of the environment. A surface model of the subject's surface is obtained, and the position and/or orientation of the probe is tracked on the surface. Image data acquired at different positons and orientations is stitched based on the tracked positon of the probe.