Aortic Segment Discrimination Using Depth Images and Deep Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for resuscitative endovascular balloon occlusion of the aorta (REBOA) require the use of X-ray fluoroscopy or ultrasound devices to identify aortic segments, which are not available in emergency situations, making it difficult to perform REBOA accurately.
Innovation Solution
A blood vessel segment discrimination system using a three-dimensional structure recognition device, depth image generation, and a deep learning model to identify aortic segments (Zone 1, Zone 2, and Zone 3) based on abdominal CT, MRI, or MRA images, enabling accurate segmentation without X-ray or ultrasound.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If X-ray fluoroscopy or ultrasound devices are used to identify aortic segments, then measurement precision and reliability are improved, but device complexity and availability worsen due to requiring specialized equipment
Solution Approach 1:
The patent replaces complex medical imaging devices (X-ray fluoroscopy, ultrasound) with a camera-based optical system that captures images and uses AI processing to identify aortic segments. This substitution maintains measurement precision while eliminating the need for specialized equipment, improving availability in emergency situations where such devices may not be accessible.
Solution Approach 2:
The system creates a digital copy of the aortic region using standard camera imaging, then processes this copy through AI algorithms to identify segments. This approach allows accurate segment identification without requiring the original complex imaging devices, enabling deployment in resource-limited emergency settings.
2Measurement precision
If deep learning models are trained using three-dimensional structure data from CT, MRI, or MRA images, then blood vessel segment discrimination accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by pre-training deep learning models using three-dimensional structure data from CT, MRI, or MRA images before actual deployment. This upfront training creates a ready-to-use model that can quickly identify aortic segments in real-time without requiring complex processing during emergency procedures, thus reducing loss of time when actually needed.
Solution Approach 2:
The patent extracts only the essential three-dimensional structural information needed for segment identification from comprehensive CT, MRI, or MRA datasets. By focusing on key geometric features rather than processing entire imaging datasets during deployment, the system maintains high discrimination accuracy while minimizing processing time and computational resource requirements.
Data Source
AI summary
A blood vessel segment discrimination system recognizes a three-dimensional structure of an abdomen of the patient, generates a depth image of the abdomen of the patient from the three-dimensional structure of the abdomen of the patient, generates a training data set used for training of a deep learning model, and discriminates an aortic segment of the patient using a trained deep learning model. A training data set generation unit generates the three-dimensional structure of the abdominal surface for learning from an abdominal CT image or the like of a person different from a patient who is a discrimination target of the aortic segment by the blood vessel segment discrimination device, generates a depth image for training from the three-dimensional structure of the abdominal surface for training, and generates a training data set showing a correspondence relationship between each pixel in the depth image for training and any of an aortic Zone 1, an aortic Zone 2, an aortic Zone 3, and another segment other than those Zones, based on the abdominal CT image.