Hearing Prosthesis Feature Extraction via Environmental Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Individuals with sensorineural hearing loss, who have limited or no residual hearing, do not benefit sufficiently from auditory prostheses that generate mechanical motion of the cochlea fluid, as they lack functional hair cells to transduce acoustic signals into nerve impulses, and those with partial sensorineural hearing loss require personalized sound processing to enhance residual hearing.
Innovation Solution
A hearing prosthesis that receives sound signals, determines the environmental classification, and performs feature extraction operations to adjust stimulation control signals based on the classification, allowing for personalized sound processing and feature-based adjustments to improve sound perception.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If auditory prostheses generate mechanical motion of the cochlea fluid, then individuals with conductive hearing loss benefit from restored hearing, but individuals with sensorineural hearing loss do not benefit sufficiently due to lack of functional hair cells
Solution Approach 1:
The system dynamically adjusts feature extraction parameters and stimulation strategies based on environmental classification and real-time signal analysis. The prosthesis transitions between different processing modes (e.g., speech-enhanced, noise-robust, music-optimized) to adapt to varying listening conditions and maximize benefit for recipients with different types of hearing loss
Solution Approach 2:
The system changes processing parameters such as feature extraction thresholds, stimulation rates, and channel allocation based on environmental context. By classifying environments (e.g., quiet, noisy, speech-heavy, music-heavy), the prosthesis optimizes its operation to provide mechanical motion when beneficial and electrical stimulation when hair cells are non-functional
2Loss of information
If feature extraction operations are performed on all sound signals, then comprehensive sound features are extracted, but processing time and computational load increase
Solution Approach 1:
The system performs partial feature extraction by selecting only the most relevant features based on environmental classification. In quiet environments, full feature sets are extracted, while in noisy environments, only noise-robust features are prioritized, reducing computational load while maintaining essential sound information
Solution Approach 2:
The system performs preliminary environmental classification and preliminary feature extraction to identify key characteristics before full processing. This preliminary analysis allows the system to pre-determine which feature extraction operations are necessary, avoiding unnecessary computational steps and reducing overall processing time
3Measurement precision
If environmental classification is determined from sound signals, then feature extraction can be optimized for the specific environment, but additional processing steps are required
Solution Approach 1:
The system merges environmental classification and feature extraction into a unified processing pipeline. Classification results directly inform feature extraction parameters in real-time, eliminating separate processing stages and reducing overall system complexity while maintaining high classification accuracy through integrated signal analysis
Data Source
AI summary
Presented herein are techniques for extracting features from sound signals received at a hearing prosthesis at least partially based on an environmental classification of the sound signals. More specifically, one or more sound signals are received at a hearing prosthesis and are converted in to stimulation control signals for use in delivering stimulation to a recipient of the hearing prosthesis. The hearing prosthesis determines an environmental classification of the sound environment associated with the one or more sound signals and is configured to use the environmental classification in the determination of a feature-based adjustment for incorporation into the stimulation control signals.