Medical Image Cutting Trajectories With Feedback Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical image segmentation methods using neural networks produce insufficiently smooth trajectories and suffer from low accuracy due to accumulated errors from discrete to continuous systems, affecting the precision of image segmentation.
Innovation Solution
A method and apparatus for generating cutting trajectories that involve selecting an initial point, performing recursive sampling with real-time deviation correction using pre-trained deep learning models, and optimizing the cutting trajectory through exponential moving averages to improve segmentation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If neural networks are used to predict displacement from image blocks, then segmentation automation is improved, but trajectory smoothness deteriorates
Solution Approach 1:
The patent implements feedback by calculating the deviation between consecutive displacement predictions and using this deviation information to adjust subsequent predictions. Specifically, the system computes the difference between the current displacement prediction and the previous displacement prediction, then uses this deviation as feedback to refine the next prediction, ensuring smoother trajectory transitions while maintaining automated segmentation.
2Productivity
If neural networks are used to learn correspondence between image blocks and displacement, then segmentation efficiency is improved, but segmentation precision deteriorates due to accumulated errors
Solution Approach 1:
The patent addresses accumulated errors by implementing a feedback mechanism that continuously monitors and corrects displacement predictions. The system calculates the deviation between consecutive predictions and uses this information to adjust subsequent predictions, preventing error accumulation and maintaining high segmentation precision throughout the automated process.
Solution Approach 2:
The patent applies preliminary action by pre-calculating and storing the correspondence between image blocks and displacement vectors during a training phase. This pre-computed knowledge is then reused during segmentation, allowing the system to make accurate predictions without performing complex calculations in real-time, thus maintaining both efficiency and precision.
3Device complexity
If direct displacement prediction from image blocks is used, then device complexity is reduced, but trajectory accuracy deteriorates
Solution Approach 1:
The patent enhances trajectory accuracy by introducing a feedback mechanism that calculates deviation between consecutive displacement predictions and uses this deviation to refine subsequent predictions. This adds only a simple computational step to the already straightforward direct prediction method, significantly improving accuracy without substantially increasing complexity.
Data Source
AI summary
The present disclosure relates to the field of computer vision technology and provides a method and apparatus for generating cutting trajectories of medical image targets. The method includes: obtaining a target image to be processed, which includes chest X-ray images, cardiac MRI images, and dermatoscope detection images; selecting an initial point of the target image and using it as the starting point for a navigation agent; and guiding the navigation agent to generate trajectory points until a cutting trajectory containing the target to be segmented is generated. Based on the generated trajectory points and the sampling areas corresponding to each sampling operation, real-time calculation is used to determine the deviation of each sampling. The method also includes correcting the sampling direction, generating cutting trajectories on the target image, and optimizing the generated cutting trajectories to obtain a target image containing the segmented target.