Application of substitution box of present cipher for automated detection of snoring sounds

作者：

Highlights：

• Classification of snoring sound in to V, O, T and E categories.

• Nonlinear feature generation function (Present-Pat) using SBox of the present cipher is proposed.

• Maximum absolute pooling decomposer is presented to obtain an effective routing method for feature generation.

• Two staged and improved version of NCA feature selector is used.

• Proposed model is tested on MPSSC dataset and attained 97.1 % accuracy and 97.6 % UAR rates.

摘要

•Classification of snoring sound in to V, O, T and E categories.•Nonlinear feature generation function (Present-Pat) using SBox of the present cipher is proposed.•Maximum absolute pooling decomposer is presented to obtain an effective routing method for feature generation.•Two staged and improved version of NCA feature selector is used.•Proposed model is tested on MPSSC dataset and attained 97.1 % accuracy and 97.6 % UAR rates.

论文关键词：Present pattern,Maximum absolute pooling,NCAINCA,Snore sound classification,Artificial intelligence

论文评审过程：Received 26 November 2020, Revised 30 April 2021, Accepted 3 May 2021, Available online 6 May 2021, Version of Record 12 May 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102085