HEART SOUNDS CLASSIFICATION WITH DEEP FEATURES AND SUPPORT VECTOR MACHINES

被引:0
作者
Demir, Fatih [1 ]
Sengur, Abdulkadir [1 ]
Cavas, Mehmet [2 ]
机构
[1] Firat Univ, Teknol Fak, Elekt & Elekt Muh Bol, Elazig, Turkey
[2] Firat Univ, Teknol Fak, Mekatronik Muh Bol, Elazig, Turkey
来源
2018 INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND DATA PROCESSING (IDAP) | 2018年
关键词
Heart Sound; Phono-cardiogram Signal; Spectrogram; Colormap; Convolutional Neural network; Support Vector Machine;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In this study, the heart which is one of the most important organ affecting life-sustaining function is examined whether it works properly in a certain rhythm or not. In this regard, an effective algorithm both analyzing and categorizing Phono-cardiogram signals (PCG) which is significant at diagnosis of diseases is presented. First of all in this context, as colored spectrogram images of heart sounds are established to be able to analyze PCG signals, the characteristic extraction maps of Convolutional Neural Networks (CNN) are used to educate the data of images obtained. CNN-VGG16 model educated previously is used when these maps are established and and it is categorized with Support Vector Machine (SVM) which is an effective classifier at machine education. The performance of all rating labels is evaluated separately for experimental study with two different data. While max performance improves about %8 in one data set (DATASETA), max performance is obtained at other dataset (DATASETB) for normal rating label.
引用
收藏
页数:5
相关论文
共 21 条
[1]   Snore Sound Classification Using Image-based Deep Spectrum Features [J].
Amiriparian, Shahin ;
Gerczuk, Maurice ;
Ottl, Sandra ;
Cummins, Nicholas ;
Freitag, Michael ;
Pugachevskiy, Sergey ;
Baird, Alice ;
Schuller, Bjoern .
18TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2017), VOLS 1-6: SITUATED INTERACTION, 2017, :3512-3516
[2]  
[Anonymous], MACH LEARN COMP ICML
[3]  
Bentley P, 2011, CLASSIFYING HEART SO
[4]   S1 and S2 Heart Sound Recognition Using Deep Neural Networks [J].
Chen, Tien-En ;
Yang, Shih-I ;
Ho, Li-Ting ;
Tsai, Kun-Hsi ;
Chen, Yu-Hsuan ;
Chang, Yun-Fan ;
Lai, Ying-Hui ;
Wang, Syu-Siang ;
Tsao, Yu ;
Wu, Chau-Chung .
IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, 2017, 64 (02) :372-380
[5]  
Deng Y., 2012, WORKSHOPCLASSIFYING
[6]   Deep Neural Networks for the Recognition and Classification of Heart Murmurs Using Neuromorphic Auditory Sensors [J].
Dominguez-Morales, Juan P. ;
Jimenez-Fernandez, Angel F. ;
Dominguez-Morales, Manuel J. ;
Jimenez-Moreno, Gabriel .
IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS, 2018, 12 (01) :24-34
[7]  
Freitag M., 2017, P INTERSPEECH STOCKH
[8]  
Gomes E. F., 2012, WORKSH CLASS HEARTSO
[9]   Classification of heart sound signal using curve fitting and fractal dimension [J].
Hamidi, Maryam ;
Ghassemian, Hassan ;
Imani, Maryam .
BIOMEDICAL SIGNAL PROCESSING AND CONTROL, 2018, 39 :351-359
[10]   Support vector machines [J].
Hearst, MA .
IEEE INTELLIGENT SYSTEMS & THEIR APPLICATIONS, 1998, 13 (04) :18-21