Detail publikace

Discrimination of Normal and Abnormal Heart Sounds Using Probability Assessment

Originální název

Discrimination of Normal and Abnormal Heart Sounds Using Probability Assessment

Anglický název

Discrimination of Normal and Abnormal Heart Sounds Using Probability Assessment

Jazyk

en

Originální abstrakt

Aims: According to the “2016 Physionet/CinC Challenge”, we propose an automated method identifying normal or abnormal phonocardiogram recordings. Method: Invalid data segments are detected (saturation, blank and noise tests) and excluded from further processing. The record is transformed into amplitude envelopes in five frequency bands. Systole duration and RR estimations are computed; 15-90 Hz amplitude envelope and systole/RR estimations are used for detection of the first and second heart sound (S1 and S2). Features from accumulated areas surrounding S1 and S2 as well as features from the whole recordings were extracted and used for training. During the training process, we collected probability and weight values of each feature in multiple ranges. For feature selection and optimization tasks, we developed C# application PROBAfind, able to generate the resultant Matlab code. Results: The method was trained with 3153 Physionet Challenge recordings (length 8-60 seconds; 6 databases). The results of the training set show the sensitivity, specificity and score of 0.93, 0.97 and 0.95, respectively. The method was evaluated on a hidden Challenge dataset with sensitivity and specificity of 0.87 and 0.83, respectively. These results led to an overall score of 0.85.

Anglický abstrakt

Aims: According to the “2016 Physionet/CinC Challenge”, we propose an automated method identifying normal or abnormal phonocardiogram recordings. Method: Invalid data segments are detected (saturation, blank and noise tests) and excluded from further processing. The record is transformed into amplitude envelopes in five frequency bands. Systole duration and RR estimations are computed; 15-90 Hz amplitude envelope and systole/RR estimations are used for detection of the first and second heart sound (S1 and S2). Features from accumulated areas surrounding S1 and S2 as well as features from the whole recordings were extracted and used for training. During the training process, we collected probability and weight values of each feature in multiple ranges. For feature selection and optimization tasks, we developed C# application PROBAfind, able to generate the resultant Matlab code. Results: The method was trained with 3153 Physionet Challenge recordings (length 8-60 seconds; 6 databases). The results of the training set show the sensitivity, specificity and score of 0.93, 0.97 and 0.95, respectively. The method was evaluated on a hidden Challenge dataset with sensitivity and specificity of 0.87 and 0.83, respectively. These results led to an overall score of 0.85.

BibTex


@inproceedings{BUT128351,
  author="Filip {Plešinger} and Juraj {Jurčo} and Pavel {Jurák} and Josef {Halámek}",
  title="Discrimination of Normal and Abnormal Heart Sounds Using Probability Assessment",
  annote="Aims: According to the “2016 Physionet/CinC Challenge”, we propose
an automated method identifying normal or abnormal
phonocardiogram recordings. Method: Invalid data segments are
detected (saturation, blank and noise tests) and excluded from further
processing. The record is transformed into amplitude envelopes in five
frequency bands. Systole duration and RR estimations are computed;
15-90 Hz amplitude envelope and systole/RR estimations are used for
detection of the first and second heart sound (S1 and S2). Features
from accumulated areas surrounding S1 and S2 as well as features
from the whole recordings were extracted and used for training.
During the training process, we collected probability and weight values
of each feature in multiple ranges. For feature selection and
optimization tasks, we developed C# application PROBAfind, able to
generate the resultant Matlab code. Results: The method was trained
with 3153 Physionet Challenge recordings (length 8-60 seconds; 6
databases). The results of the training set show the sensitivity,
specificity and score of 0.93, 0.97 and 0.95, respectively. The method
was evaluated on a hidden Challenge dataset with sensitivity and
specificity of 0.87 and 0.83, respectively. These results led to an
overall score of 0.85.",
  booktitle="Computing in Cardiology",
  chapter="128351",
  edition="2016",
  howpublished="print",
  year="2016",
  month="september",
  pages="1--1",
  type="conference paper"
}