Search

Health is one of the aspects of life that the emerging technology has changed. Automated systems, powered by machine learning, can analyze large volumes of data of many forms, text, image, or audio. The newer automated systems still do not replace the medics, they are just instruments. A mixed approach can be applied where the patient uses an application to get a verdict, but he/she consults a medic to find out more or to get the necessary medication. In our research years, we discovered how machine learning is used in medicine by studying the existing works, and proposed other solutions. Even if they have aspects in common with other types of applications, medical applications require a specific, stricter approach (correctness, usability, etc.). In our projects, we were making particular engineering decisions by considering the patient's interest - do we need to store the recordings?, when should the medic be involved in the medical analysis? , etc. We also noticed that even if the machine learning techniques are general, approaching different diseases requires us to adapt our methods. This paper does not cover our research fully, but we consider it is an interesting work summarizing the good practice principles in medical software and showing our own works as applications. [ABSTRACT FROM AUTHOR]

The technological progress in computer science (particularly, in machine learning) has contributed to the improvement of medical services, both in detecting and treating diseases. The large volumes of data, that are overwhelming for human experts (doctors, nurses), can easily be managed by automated systems, as long as we have the computational resources. Obviously, human experts are still essential in the process - we think of the use of computer science in medicine as a collaboration between medical staff and artificial intelligence. The usual types of data that can be processed by automated systems are text, sound, and image types. In this paper, we approach the diagnosis subject and focus on data consisting of sound. We created a heart murmur detection system - it analyzes recordings and tells the user whether the sound samples indicate a heart murmur or not, based on a trained machine learning model. One of the main advantages of our system is the fact that we ran a large number of experiments, with different configurations of denoising techniques and features taken into consideration. We were able to draw some interesting conclusions, for example, we found out which features are the most important for the classification and which features are not worth computing. Also, our work denotes a thorough understanding of sound processing. [ABSTRACT FROM AUTHOR]