Abstract

Info

Title:: Experimental study on detection of common bacteria in wound infection by electronic nose

Author(s):: LU Yanyi; ZENG Lin; YAN Bowen; LI Min; HE Qinghua; State Key Laboratory of Trauma, Burns and Combined Injury/Daping Hospital, Army Medical University, Chongqing 400042, China

Keywords:: Keywords: bacteria in wound infection electronic nose classification and recognition

Abstract:: Abstract: To further explore the application of electronic nose in the rapid screening of bacterial infection of wounds, a self-made electronic nose is used to detect Thioglycolate (TH) broth of Acinetobacter baumannii (A. Baumannii), Escherichia coli (E. Coli), Klebsiella pneumoniae (K. pneumoniae), Staphylococcus aureus (S. aureus), and Pseudomonas aeruginosa (P. aeruginosa) and sterile TH broth. After preprocessing, the data are classified by support vector machine, BP neural network and logistic regression, separately. The results show that BP neural network can achieve an overall recognition rate over 90%, and specific recognition rates over 98% for E. Coli, K. pneumoniae and S. aureus, and that its recognition rates for A. Baumannii and P. aeruginosa are around 90%, indicating that electronic nose is feasible for the rapid detection and recognition of bacteria in wound infection.