Combining sparse coding and time-domain features for heart sound classification

View Full Paper

Ask AI

Bookmark

Population

Audio PCG data from the 2016 PhysioNet/CinC Challenge

Comparison

Modified sparse coding combined with time-domain… vs Original sparse coding algorithm submitted to…

Design

Other

Authors

Discussion

Combining sparse coding with time-domain features may enhance automated heart sound classification; leaves open whether.

Population

Audio PCG data from the 2016 PhysioNet/CinC Challenge

Intervention

Modified sparse coding (using a matrix norm in the dictionary update step) combined with time-domain features using support vector machines (SVMs)

Comparator

Original sparse coding algorithm submitted to the challenge

Outcome

Cross-validated mean accuracy (MAcc) for heart sound classificationsurrogate

Combining sparse coding with time-domain features improves the accuracy of automated heart sound classification.

Whitaker et al. (2017) studied this question.

View Full Paper

Ask AI

Bookmark