Cochlear Implant Harmonic Frequency Estimation Using Active Contours
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cochlear implant systems face challenges in accurately estimating instantaneous frequencies, especially in the presence of multiple harmonics and noise, leading to poor speech intelligibility and pitch perception due to broad filter bank resolutions and limited electrode channels.
Innovation Solution
The implementation of an active contour model to estimate dominant frequencies from a spectrogram generated by short-time Fourier transformation, allowing for channel-specific sampling sequences and improved spatial information delivery to electrode contacts in the cochlear implant array.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional filter bank methods are used for frequency estimation, then the system is simple to implement, but the frequency estimation precision deteriorates in the presence of multiple harmonics and noise
Solution Approach 1:
The patent introduces an active contour model as an intermediary computational framework between the raw spectrogram and the final frequency estimates. This model acts as a mediator that iteratively refines frequency predictions by minimizing an energy functional, thereby achieving superior frequency estimation precision in noisy conditions with multiple harmonics compared to traditional filter bank methods
Solution Approach 2:
The system performs preliminary action by generating a spectrogram through short-time Fourier transformation before applying the active contour model. This preliminary spectral analysis provides the foundation for subsequent iterative frequency estimation, allowing the system to prepare frequency predictions in advance for each time frame
2Device complexity
If broad frequency bands are assigned to each electrode channel, then the device complexity is reduced, but the speech intelligibility and pitch perception deteriorate
Solution Approach 1:
The patent applies local quality by assigning different center frequencies to each electrode channel and using channel-specific sampling sequences tailored to each channel's frequency characteristics. This allows each electrode contact to be optimized for its specific frequency region, improving speech intelligibility and pitch perception without increasing the number of physical electrodes
Solution Approach 2:
The system implements dynamics by using time-varying stimulation rates that adapt to the instantaneous frequency content of each channel. The stimulation pulse rate varies dynamically based on the detected frequency, allowing the system to respond flexibly to changing speech signals while maintaining a relatively small number of electrode channels
3Ease of operation
If constant stimulation rate is applied across all electrode channels, then the device operation is simplified, but the pitch perception and speech intelligibility deteriorate
Solution Approach 1:
The patent implements local quality by applying channel-specific sampling sequences where each electrode channel operates at a stimulation rate tailored to its center frequency and the detected instantaneous frequency. This allows different parts of the frequency spectrum to be processed with appropriate temporal resolution, preserving pitch information while maintaining operational simplicity through automated frequency tracking
Solution Approach 2:
The system performs preliminary action by estimating the instantaneous frequency of each channel before generating the stimulation pattern. This preliminary frequency estimation allows the system to pre-determine the appropriate stimulation rate for each channel, simplifying the overall operation while improving pitch perception
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A signal processing arrangement generates electrical stimulation signals to electrode contacts in an implanted cochlear implant array. An input sound signal is processed to generate band pass signals that each represent an associated band of audio frequencies. A spectrogram representative of frequency spectrum present in the input sound signal is generated. A characteristic envelope signal is produced for each band pass signal based on its amplitude. An active contour model is applied to estimate dominant frequencies present in the spectrogram, and the estimate is used to generate stimulation timing signals for the input sound signal. The electrode stimulation signals are produced for each electrode contact based on the envelope signals and the stimulation timing signals.