Pulse Rate Identification Using Wavelet and Spectral Transforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current pulse oximetry systems face challenges in accurately measuring pulse rates due to noise and artifacts in photoplethysmographic (PPG) signals, which can lead to inaccurate oxygen saturation and pulse rate calculations.
Innovation Solution
The use of continuous wavelet transforms (CWT) and spectral transforms in conjunction with Fourier transforms to identify and filter pulse rates, combining results through averaging or selecting the closest value to historical data, to improve accuracy and ignore temporally discrete artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If spectral averaging techniques are used to calculate pulse rate, then computational efficiency is improved, but measurement precision deteriorates due to inability to track temporal changes and resolve discrete artifacts
Solution Approach 1:
The patent segments the pulse rate calculation process into two distinct stages: first using spectral transform for efficient frequency region identification, then applying CWT for precise pulse rate determination. This segmentation allows each method to operate in its optimal performance zone, with spectral methods handling the initial screening and CWT handling the final precise measurement.
Solution Approach 2:
The patent introduces dynamic adaptability by using spectral transform results to dynamically guide the CWT process. The spectral analysis identifies frequency regions of interest that become the focal point for subsequent CWT analysis, allowing the system to adapt its computational resources to the most relevant signal components rather than uniformly processing all frequencies.
2Measurement precision
If CWT is used to track pulse rate changes through time, then measurement precision is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary spectral analysis before applying CWT to identify frequency regions of interest. This preliminary action filters out irrelevant frequency components, allowing the computationally intensive CWT to focus only on the most promising frequency ranges, thereby reducing the overall computational burden.
Solution Approach 2:
The patent applies CWT selectively at specific scales corresponding to identified frequency regions of interest rather than uniformly across all frequencies. This local application of CWT concentrates computational resources on the most relevant signal components, improving efficiency while maintaining measurement precision.
3Productivity
If spectral transform alone is used to identify pulse rate, then computational efficiency is maintained, but reliability deteriorates due to temporal resolution limitations
Solution Approach 1:
The patent merges the strengths of two different transform methods by combining spectral transform and CWT in a unified pulse rate identification system. The spectral transform provides efficient frequency domain analysis while CWT provides time-scale analysis, and their results are integrated to produce a more reliable pulse rate measurement than either method alone.
Solution Approach 2:
The patent uses spectral transform results as an intermediary to guide the CWT process. The frequency regions identified by spectral analysis serve as intermediate targets that focus the subsequent CWT analysis, creating a bridge between the efficient but less precise spectral method and the precise but computationally intensive CWT method.
Data Source
AI summary
According to embodiments, techniques for using continuous wavelet transforms and spectral transforms to identify pulse rates from a photoplethysmographic (PPG) signal are disclosed. According to embodiments, candidate pulse rates of the PPG signal may be identified from a wavelet transformed PPG signal and a spectral transformed PPG signal. A pulse rate may be determined from the candidate pulse rates by selecting one of the candidate pulse rates or by combining the candidate pulse rates. According to embodiments, a spectral transform of a PPG signal may be performed to identify a frequency region associated with a pulse rate of the PPG signal. A continuous wavelet transform of the PPG signal at a scale corresponding to the identified frequency region may be performed to determine a pulse rate from the wavelet transformed signal.