Soft Tissue Emulation Using Non-Autoregressive Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current soft tissue emulation systems for medical applications, particularly for cardiac and hepatic tissues, face challenges in achieving real-time simulation due to the limitations of numerical solvers in speed and accuracy, which hinders their translation into clinical environments.
Innovation Solution
A computer-implemented soft tissue emulation system utilizing an artificial neural network (ANN) with non-autoregressive inference to generate a digital twin of soft tissue, allowing for real-time simulation of tissue responses to thermal and electromechanical stimuli, independent of previous states, and enabling fast and accurate emulation of cardiac and hepatic tissue behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional numerical solvers are used to simulate soft tissue behavior, then the simulation accuracy is maintained, but the simulation speed is too slow for real-time clinical applications
Solution Approach 1:
The patent creates a digital twin (copy) of the soft tissue that mimics the behavior of the original tissue. This digital twin is generated using an artificial neural network trained on data from traditional biomechanical models, allowing real-time simulation without repeatedly running slow numerical solvers. The copy enables fast prediction of tissue response to stimuli while maintaining accuracy through the trained network.
Solution Approach 2:
The artificial neural network is trained in advance using comprehensive biomechanical modeling data. This preliminary training phase allows the network to learn complex tissue behaviors beforehand, so that during actual clinical use, the pre-trained network can rapidly predict tissue responses without requiring real-time numerical solving, thus achieving both speed and accuracy.
2Productivity
If traditional biomechanical models are used for soft tissue simulation, then the physiological and physical insights are incorporated, but the computational complexity and time requirements prevent clinical translation
Solution Approach 1:
The patent replaces the traditional mechanical numerical solving system with an artificial neural network-based system. The complex biomechanical models that require extensive computational resources are substituted with a trained neural network that can perform predictions rapidly, reducing computational complexity while maintaining the physiological and physical insights embedded in the training data.
3Speed
If real-time simulation of soft tissue is achieved using AI methods, then the speed requirement is met, but the current state-of-the-art AI implementations are still far from real-time performance
Solution Approach 1:
The patent segments the simulation process into two distinct phases: an offline training phase where the neural network is trained using comprehensive biomechanical model data, and an online inference phase where the pre-trained network rapidly predicts tissue responses. This segmentation allows computationally intensive work to be done beforehand, enabling real-time performance during actual use without sacrificing accuracy.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
A soft tissue emulation system, comprising: an input interface, configured to obtain imaging data of the soft tissue; a computing unit, configured to implement an artificial neural network, which is adapted to generate, using the obtained imaging data as input, and a biophysical model of the soft tissue, a digital twin of the soft tissue at different times, wherein the biophysical model describes the response of the soft tissue to thermal and/or electromechanical stimuli over time, and wherein the generation of the digital twin at one time is independent of the generation of the digital twin at another time; and an output interface, configured to output a representation of the soft tissue over time based on the digital twin generated by the artificial neural network.