Collecting Lung Sound Signals to Track Air Pollution and Teach Machine Learning Algorithms

In 2010, Julio and his team started collecting lung sound signals from students at the Instituto Tecnologico de Mexicali (ITM). They were concerned about asthma and air pollution in the region. The students’ ages ranged from 18 to 26 years old and included both males and females. 12 years later, the work continues. Julio recently won IEEE DataPort’s data upload contest in the biomedical category for his research.

How the Datasets are Used Today

Julio published his research titled “Cardiopulmonary Sounds Database” on IEEE DataPort along with corresponding articles on IEEE Xplore. The datasets can be applied to intelligent systems in machine learning and deep learning to recognize and classify normal signals of lung sounds from healthy subjects.

The datasets have also been used in engineering courses. The signals are used in courses of digital signal processing to analyze the signals in frequency and time, applying IIR and FIR filters to suppress noise. The signals are also used in research projects to reinforce other datasets researchers have collected.

Benefits of IEEE DataPort

“IEEE DataPort is wonderful. It gave us an excellent opportunity to discover new topics and data.”

Julio said he uses IEEE DataPort for the access to large datasets of diverse topics which he said is useful to start research or develop a class exercise.

“IEEE DataPort helped us meet our research goals since it can be shared with other researchers, teachers, and professors. In machine learning and deep learning, it is vital to count as many signals as possible to achieve an accurate result.”

See the dataset, learn more about IEEE DataPort, or upload your own research data.