Visual Image Processing for Irregular Phosphene Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current visual prostheses and image processing systems face challenges with irregular phosphene patterns due to limited spatial resolution, dynamic range, and unpredictability, requiring complex calibration processes and ongoing adjustments, which are also applicable to other irregular output image systems like VR and AR.
Innovation Solution
A visual image processing system that decouples sensor maps from visual percepts, allowing flexible and computationally efficient processing by using digital image sensors, depth sensors, or accelerometers to generate stimulus control information based on predefined regions and threshold values, enabling adaptable mapping between stimuli and visual percepts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a training/calibration process is used to map stimuli to phosphenes, then the accuracy of phosphene mapping is improved, but the time required for system setup and ongoing adjustments increases
Solution Approach 1:
The system performs preliminary mapping of sensor data regions to phosphene patterns during system setup, creating a lookup table that stores pre-determined stimulus configurations. During operation, the system directly queries this pre-computed mapping rather than performing real-time calibration, significantly reducing operational time while maintaining mapping accuracy.
Solution Approach 2:
The system creates a simplified computational model that replicates the complex stimulus-phosphene relationships in a condensed format. This model copy allows the system to quickly determine appropriate stimuli without repeatedly performing full calibration procedures, reducing time loss while preserving mapping fidelity.
2Adaptability or versatility
If complex processing is used to handle irregular phosphene patterns, then the adaptability of the system is improved, but the computational complexity increases
Solution Approach 1:
The system divides the visual field into discrete sensor regions, each independently mapped to specific phosphene patterns. This segmentation allows irregular patterns to be handled as composed elements rather than requiring complex holistic processing, reducing computational complexity while maintaining adaptability to irregular configurations.
Solution Approach 2:
The system transforms the representation of visual data by changing parameters from continuous irregular coordinates to discrete region-based indexing. This parameter transformation simplifies the mathematical operations required to handle irregular phosphene patterns, reducing processing complexity while preserving the ability to adapt to various pattern configurations.
3Measurement precision
If iterative refinement with multiple measurement cycles is used, then the phosphene map accuracy is improved, but the productivity of the system decreases
Solution Approach 1:
The system performs comprehensive mapping measurements during an initial setup phase, storing the results in a pre-computed lookup table. This preliminary action consolidates multiple measurement cycles into a single setup period, improving phosphene map accuracy while preventing time loss during subsequent operational cycles.
Solution Approach 2:
The system anticipates future measurement needs by pre-computing and storing multiple stimulus-phosphene mappings in advance. This cushioning approach ensures that even if iterative refinement is needed, the computational work has already been performed, protecting productivity during actual operation while maintaining measurement precision.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A visual image processing method is based on received spatial field information (302) from a spatial field sensor (116). A data store (804) is accessed, which contains a sensor map data structure (206) comprising a set of predefined regions (208) within a spatial field corresponding with the information received via the sensor input. Each predefined region is associated in the data structure with one or more of a set of stimuli (204) applicable to a biological visual system, and each stimulus corresponds with a visual percept (210). The spatial field information associated with each region is processed to generate stimulus control information which is applied to select, from within the sensor map data structure, stimuli from the set of stimuli for application to the biological visual system. Output stimulus signals (310) are generated, which are suitable for application to the biological visual system based upon the selected stimuli. Flexible mappings are thus provided between visual percepts, stimuli which may be applied (e.g. via a prosthetic implant) in order to generate the percepts, and associations between those stimuli and regions of the spatial field corresponding with the visual percepts.