A simple lung sound enhancement for automatic identification of lung pathologies

College

College of Computer Studies

Department/Unit

Computer Technology

Document Type

Archival Material/Manuscript

Abstract

Auscultation or lung sound analysis depends on the familiarity of the physician on detecting sound patterns. However, typical environment for auscultation are performed in rooms susceptible to different sounds such as vocal sound, ventilation machines and ambient noise, which may impede the subjective evaluation of the lung sounds. This paper presents a simple signal enhancement scheme for normal lung sounds in order to successfully extract the features which include the bandwidth, peak frequency and center frequency. The extracted features could be used in automatic detection and classifications of lung sound abnormalities of different. Results show that the enhanced signal has features in the 300 to 700 Hz range while the raw and denoised signals have features below 300 Hz. Listening test shows improved score in enhanced signals over scores on the raw and denoised signals with an average score of 1.3 over 1.03 in raw and 0.82 in denoised signals.

html

Disciplines

Computer Sciences

Note

Presented at the AUN/SEED-Net Regional Conference in Electrical and Electronics Engineering International Symposium on Multimedia and Communication Technology (ISMAC) 2013, ASEAN Energy Focus, February 4-5, 2013, Holiday Inn Bangkok Hotel, Ratchaprasong, Bangkok, Thailand.

Keywords

Lungs—Sounds; Auscultation; Noise control—Equipment and supplies

Upload File

wf_no

Share

COinS