Artificial Vision Parameter Learning for Phosphene Image Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional human psychophysical experiments for evaluating visual prosthetic systems are time-consuming, costly, and prone to subject variability, making it difficult to gather statistically significant data, especially during pandemics or other gathering restrictions.
Innovation Solution
An artificial vision parameter automating method using a sensory module, image processing, learning, and storing modules to generate low-resolution phosphene images and machine learning models that predict clinical results, reducing the need for extensive human testing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human psychophysical experiments are conducted to evaluate visual prosthetic systems, then clinical results can be obtained, but the process becomes time-consuming and costly
Solution Approach 1:
The patent creates a computational model that copies and simulates human visual processing and psychophysical responses. Instead of directly testing human subjects for every parameter evaluation, the system uses a trained neural network model that replicates human visual cortex responses to phosphene images, enabling rapid prediction of clinical outcomes without repeated human experiments
Solution Approach 2:
The system performs preliminary training of the computational model using a dataset of phosphene images and corresponding human psychophysical responses. This preliminary action creates a pre-trained model that can predict clinical results for new visual prosthetic configurations without requiring new human subjects, thus saving time in the evaluation process
2Reliability
If a large number of test subjects are recruited to gather statistically significant data, then measurement reliability improves, but costs and time requirements increase
Solution Approach 1:
The computational model serves as a virtual copy of human visual processing that can be tested repeatedly with different parameters without recruiting additional human subjects. The model maintains statistical reliability by being trained on a comprehensive dataset that captures variability in human responses, allowing extensive parameter exploration with a single trained model
Solution Approach 2:
The system uses itself (the trained computational model) to evaluate multiple test cases and generate statistically significant results without requiring external human subjects for each test. The model performs self-evaluation across numerous configurations, maintaining reliability while reducing the quantity of human subjects needed
3Manufacturing precision
If psychophysical tests are performed repeatedly to evaluate different visual prosthetic configurations, then system performance can be optimized, but subject fatigue and concentration loss occur
Solution Approach 1:
The patent replaces repeated human subject testing with computational model evaluation. The model can assess numerous visual prosthetic configurations without experiencing fatigue or concentration loss, enabling thorough system optimization while eliminating the operational difficulties associated with human subject sustainability
4Measurement precision
If human psychophysical experiments are conducted during pandemics or gathering restrictions, then clinical evaluation can proceed, but recruitment and execution become difficult
Solution Approach 1:
The computational model provides a virtual alternative to in-person human subject experiments. During pandemics or gathering restrictions, the pre-trained model can evaluate visual prosthetic systems remotely without requiring subject recruitment, transportation, or physical testing facilities, thus maintaining clinical evaluation capability while adapting to restricted environments
Data Source
AI summary
In artificial vision parameter automating method, the system includes a sensory module, a communication module, a database, an image processing module, a learning module and a automating and storing module. The sensory module is configured to take images. The communication module is configured to be send the data obtained from the sensory module to the database. The database is configured to store the images received from the communication module. The image processing module is configured to convert a selected image from the database into a low-resolution phosphene image, to generate test images with various artificial vision parameters. The learning module is configured to extract important features from high-resolution training images, and to build various machine learning models competent at identifying visual stimuli from low-resolution phosphene images. The automating and storing module is configured to automatically select and store primary test data, based on the learning in the learning module.