Cardiac Functional Indices from Fundus Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining cardiac functional indices and predicting adverse cardiovascular events rely heavily on expensive and inaccessible cardiovascular imaging technologies, limiting their availability, especially in developing countries.
Innovation Solution
A computer-implemented method that uses a captured fundus image to estimate cardiac functional indices by encoding the image into a joint latent space, decoding it to represent the patient's heart, and inputting this representation into a neural network to generate the desired indices, thereby avoiding the need for direct heart imaging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If cardiovascular imaging tests (CMR, CT) are used to determine cardiac functional indices, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent creates a virtual copy of cardiovascular imaging data by training a neural network to generate CMR or CT images from retinal fundus images. The generated images replicate the appearance and diagnostic information of actual cardiovascular scans without requiring the expensive imaging equipment, thus resolving the contradiction between measurement precision and device complexity
Solution Approach 2:
The patent introduces retinal fundus images as an intermediary medium that bridges the gap between easily accessible primary care data and the information needed for accurate cardiac functional assessment. The neural network acts as a mediator that translates retinal image information into cardiovascular diagnostic data, eliminating the need for direct cardiovascular imaging
2Reliability
If cardiovascular imaging tests are used to assess patient risk, then reliability is improved, but accessibility deteriorates
Solution Approach 1:
The system copies the diagnostic capability of cardiovascular imaging into a computational model that can run on standard computers in primary care settings. By generating synthetic cardiovascular images from retinal photos, the system makes reliable CVD risk assessment accessible without requiring specialized imaging equipment
Solution Approach 2:
Instead of capturing images of the heart directly (the conventional approach), the patent inverts the process by capturing images of the retina and using AI to infer cardiovascular information. This inversion enables reliable cardiac assessment through equipment already available in primary care optical clinics
3Ease of operation
If retinal images are used instead of cardiovascular images, then ease of operation is improved, but measurement precision deteriorates
Solution Approach 1:
The patent replaces the mechanical imaging system (CMR or CT scanners) with an information processing system (neural network). The neural network processes retinal images to generate synthetic cardiovascular images, substituting physical imaging mechanics with computational inference while maintaining measurement precision
Solution Approach 2:
The system changes the parameter space by training the neural network on paired datasets of retinal images and corresponding cardiovascular images. This parameter transformation enables the network to learn the mapping between retinal features and cardiac functional indices, achieving accurate measurements from easily accessible images
Data Source
AI summary
A computer-implemented method for determining cardiac functional indices for a patient including: receiving an image of a fundus of the patient; encoding the received image into a joint latent space; decoding from the joint latent space a representation of the patient's heart; providing the representation decoded from the joint latent space to a neural network configured to generate cardiac functional indices; and outputting the cardiac functional indices generated by the neural network in response to receiving the decoded representation of the patient's heart.