Feature Analysis of Snore Signals and Other Sound Signals Based on Complex Order Derivative Processing

The key to snore recognition is to process snore signals and acquire snore features. There is often interference from other sounds during sleep, such as cough sounds and tapping sounds. The snore features extracted by the existing snore processing methods are not significantly different from the environmental noise during sleep, which leads to the complicated algorithm and low accuracy of snore recognition when there is environmental noise interference. In this paper, the complex order derivative was used to process snore signals and extract snore features. The experimental results showed that the snore signals processed by the complex order derivative were obviously different from the ambient noise, and it can be applied to snore recognition.