Published Papers

Description: In recent years there has been a rise in the number of patients suffering from acute and chronic pulmonary diseases because of varying reasons like pollution, lung damage, or infections. The following research is regarding a Neural Network based solution for the recognition of the abnormality and possible disease based on lung auscultation. The following paper depicts that RNN-LSTM and CNN were the best-performing techniques. Although a higher percentage of noise while capturing the auscultation audio and limited data leads to a saturation point for the models to improve. The dataset had over 5000 breathing cycles for COPD, whereas only about 100 breathing cycles for LRTI and URTI. This unbalanced data made it difficult for the models to perform well on test audio clips because of the bias introduced by the large count of COPD samples. We adopted a filter-based audio augmentation to rebalance the dataset. To get the most out of the data we had, we utilized multiple features like MFCC, Chromagram, and Spectrogram extracted from the same audio clip. Since these extracted features are not fathomable to humans, we used convolutional neural networks to perform primary feature extraction. Likewise, dedicated CNN models acted as feature extractors whereas the dense neural network served as the actual classifier. We developed multiple versions of the models with fine-tuned parameters. The ML models based on a single feature were considered the benchmark for evaluating more complex, multi-feature DL models.

Description: The field of medical science is getting more effective with emerging trends of computer science technologies like AI, ML, and DL. The area where these technologies play an important role is the detection or recognition of diseases. This paper discusses existing methodologies and various steps involved in the process of detection/recognition of pulmonary diseases using lung sound. The paper is divided into 4 sections. These sections include general steps of any recognition system using lung sound and study of existing methods. The paper helps to overview the different approaches/experiments and build new and effective ones which can give better accuracy.