The use of machine learning methods can facilitate the work of doctors, for example, the current research aims to speed up the analysis of smears on cervical cancer screening using artificial intelligence – revealed the University of Debrecen, the Széchenyi István University of Győr and the Hungarian Industrial Innovation Mathematical Services (HU-MATHS-IN) for business organizations on Wednesday.
The communication quotes Balázs Harangi, an assistant professor in the Department of Computer Graphics and Image Processing at the University of Debrecen, who said he had previously conducted research to classify eye diseases due to diabetes through the pupil using traditional learning algorithms.
As he said, a large number of digital images can be compiled into an expert-labeled database through which “deep-learning networks” learn to distinguish between images and may even recognize discriminatory visual features that even doctors are not tied to the disease when examined visually. During the learning process, they optimize themselves to achieve the highest possible accuracy, the assistant professor said. An example of this is a project to develop a mobile phone application for laypeople that can diagnose a skin lesion, whether it is abnormal, and whether it is worth seeing a dermatologist as soon as possible, he added.
Reporting from ongoing research, he said the project is about using automated learning tools to automatically analyze cytological findings from cervical cancer screening, which can increase accuracy and most importantly speed.
According to the announcement, Zoltán Horváth, president of HU-MATHS-IN and the president of the European organization, EU-MATHS-IN, professor of Széchenyi István University, said at the press conference that industrial mathematics is the result of mathematical research (modeling, simulation, optimization, industrial, economic and, in particular, innovation. Its applications are extremely wide, covering almost all sectors, from the automotive industry to agriculture and the healthcare industry, which promises “forint-worthy” benefits for economic actors.
He added that industrial mathematics research groups achieve important results by bringing together several disciplines and adding relevant mathematical content.