School of Computing

Jan 22
14:30 - 15:00
Data Science: Lam Pham
Data Science Group Seminar
Deep Learning Frameworks Applied For Respiratory Disease Detection

Abstracts: This work presents deep learning frameworks applied for lung sound classification, aiming to detect respiratory diseases. In particular, a whole detecting process comprises of two main steps. Firstly, respiratory sounds are transformed into spectrogram representation, showing both spectral and temporal information. This step is referred to front-end feature extraction. Next, these spectrograms are fed into a back-end deep learning model for classification. In this work, experiments are conducted over a benchmark dataset of respiratory sounds, namely ICBHI. Obtained results show two main contribution: Firstly, this work provides an intensive analysis of factors such as respiratory cycle length, time resolution, and network architecture, which affect final predictions. Secondly, proposed deep learning frameworks for detecting respiratory diseases outperform the state of the art..

Location

Intellectual Hub,
M3-27,
Medway Building,
University of Kent,
Chatham Maritime,
Kent,
ME4 4AG
United Kingdom
Map

Details

Open to all,

Contact: Daniel Soria
E: D.Soria@kent.ac.uk
School of Computing

School of Computing, University of Kent, Canterbury, Kent, CT2 7NF

Enquiries: +44 (0)1227 824180 or contact us.

Last Updated: 14/08/2015