Simulated Intraoperative Fluorescence Imaging Without Contrast Agents
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fluorescence imaging is underutilized in surgery due to the inconvenience of administering imaging agents, limiting its practical application in intraoperative imaging.
Innovation Solution
A system that generates simulated intraoperative fluorescence images using a trained generative adversarial network (GAN) model based on intraoperative white light images, eliminating the need for imaging agents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fluorescence imaging is used to improve surgical accuracy and visualization, then anatomical structures can be better visualized, but the process becomes inconvenient and impractical due to requiring imaging agent administration and waiting time
Solution Approach 1:
The system creates a simulated fluorescence image as a copy of what would be observed with actual fluorescence imaging, using a trained GAN model to translate white light images into fluorescence-like images. This copying approach allows surgeons to obtain the visualization benefits of fluorescence imaging without the complexity of agent administration and waiting time
Solution Approach 2:
The patent replaces the mechanical and chemical system of fluorescence imaging (involving imaging agent injection, circulation, and fluorescence detection) with a computational system using machine learning models. The GAN-based translator substitutes the physical fluorescence mechanism with an algorithmic image transformation process, eliminating the need for imaging agents while preserving the visualization benefits
2Loss of information
If imaging agents are administered to enable fluorescence imaging, then anatomical structures become visible, but surgical time is increased due to waiting for agent distribution
Solution Approach 1:
The GAN model is trained in advance on paired white light and fluorescence images to learn the transformation relationship. During surgery, this pre-trained model can instantly generate fluorescence-like images from white light images without requiring the time-consuming process of imaging agent administration and waiting for distribution throughout the body
Solution Approach 2:
Instead of waiting for the natural fluorescence imaging process to complete (which requires time for agent circulation and distribution), the system creates an immediate computational copy of the expected fluorescence image appearance, providing the same information about anatomical structures without the time delay
Data Source
AI summary
The present disclosure relates generally to medical imaging, and more specifically to machine-learning techniques to generate intraoperative fluorescence images of a subject (e.g., to aid a surgery, to aid diagnosis and treatment of diseases). The system can receive an intraoperative white light image of the subject, input the intraoperative white light image of the subject into a generator of a trained generative adversarial network (GAN) model trained. In some examples, the GAN model is trained using a plurality of training image pairs, and each training image pair comprises an intraoperative white light training image and an intraoperative fluorescence training image of a same tissue. The system can obtain, from the generator, the generated intraoperative fluorescence image of the subject and display, on a display, the generated intraoperative fluorescence image of the subject.