Recognition of pulmonary diseases from lung sounds using convolutional neural networks and long short-term memory

In this paper, a study is conducted to explore the ability of deep learning in recognizing pulmonary diseases from electronically recorded lung sounds. The selected data-set included a total of 103 patients obtained from locally recorded stethoscope lung sounds acquired at King Abdullah University Hospital, Jordan University of Science and Technology, Jordan. In addition, 110 patients data were added to the data-set from the Int. Conf. on Biomedical Health Informatics publicly available challenge database. Initially, all signals were checked to have a sampling frequency of 4 kHz and segmented into 5 s segments. Then, several preprocessing steps were undertaken to ensure smoother and less noisy signals. These steps included wavelet smoothing, displacement artifact removal, and z-score normalization. The deep learning network architecture consisted of two stages; convolutional neural networks and bidirectional long short-term memory units. The training of the model was evaluated based on a k-fold cross-validation scheme of tenfolds using several performance evaluation metrics including Cohen's kappa, accuracy, sensitivity, specificity, precision, and F1-score. The developed algorithm achieved the highest average accuracy of 99.62% with a precision of 98.85% in classifying patients based on the pulmonary disease types using CNN + BDLSTM. Furthermore, a total agreement of 98.26% was obtained between the predictions and original classes within the training scheme. This study paves the way towards implementing deep learning models in clinical settings to assist clinicians in decision making related to the recognition of pulmonary diseases.

SUPPLEMENTARY INFORMATION

The online version supplementary material available at 10.1007/s12652-021-03184-y.

Insights into pathological and molecular characterization of avipoxviruses circulating in Egypt

1. Avipoxvirus (APV) infections are one of many threats inflicting economic losses within the poultry industry, particularly in tropical and subtropical countries. A proper and comprehensive study for APVs is needed to increase the knowledge concerning the diversity and evolution of the virus.2. For this purpose, 136 bird flocks of different species and breeding types were examined for APV infection between October 2016 and November 2017. One hundred and thirty samples had visible pocks on the chorioallantoic membrane (CAM) which were designated as fowl pox-like viruses via amplification of 578 bp from the P4b gene and 1800 bp from the fpv140 locus.4. A comprehensive phylogenetic analysis of fpv167 locus (P4b), fpv140 locus (fpv139 and fpv140) and fpv94 (DNA polymerase) revealed that all the analysed strains belong to fowl pox-like viruses (clade A; subclade A1 and A2). Based on the fpv140 locus full nucleotide sequence, three turkey originated strains were seen to be divergent from chicken originated sequences and branched into novel subclade A1.b.5. Trees comparison, within the term of speculation of virus-host specificity, clearly highlighted a high order specific subgrouping among subclades in the case of the fpv140 locus (including fpv139 and fpv140). Hence, the fowl poxvirus, turkey poxvirus and pigeon poxvirus strains clustered into distinct host-specific subclades A1a, A1.b and A2, respectively, which could not be seen in the FWPV-P4b and DNA polymerase phylogeny.