Congratulations to Shengsheng Cai, Co-founder of Minttihealth! He has received an acceptance letter from Panicos Kyriacou, BESc,PhD, Editor in Chief of Biomedical Signal Processing and Control, acknowledging him as a co-author of a paper titled “Cardiopulmonary auscultation enhancement with a two-stage noise cancellation approach.” The paper has been accepted for publication in the same journal as a research article.
The paper details the development of a two-stage approach for denoising cardiopulmonary sounds. In the first stage, a conventional ANC method was utilized for coarse noise cancellation, while the second stage employed the proposed DINC-Net, which used dual-inputs of the coarsely denoised auscultation signal and the estimated interference from the first stage.
The DINC-Net consisted of two deep encoders to extract features, one interactive denoising module to generate denoising masks by exploiting mutual information, and one deep decoder to produce the final denoised output.
Simulation experiments and real auscultation applications demonstrated that the proposed two-stage method outperformed both one-stage methods and two-stage methods that did not incorporate the interactive denoising module.
The study suggests that the proposed two-stage approach could be a promising technical route for noise cancellation in real online auscultation applications of electronic stethoscopes.
The research was supported by the Suzhou Science and Technology Project, and co-author Shengsheng Cai, Co-founder of Minttihealth, was responsible for validation, investigation, and data curation during the research.