Posters

Cardiotocography and ultrasound in obstetrical practice

Ariana Rabac

Monitoring the fetus during pregnancy and childbirth is a great challenge for the obstetric team. Proper monitoring of the fetus during pregnancy and childbirth is very important in order to detect deviations and complications in a timely manner and thus for a better perinatal outcome for the newborn and the mother. In this paper, a brief description of cardiotocography and ultrasound, as two main methods for fetus monitoring is provided.

Use of Regressive Artificial Intelligence and Machine Learning Methods in Modelling of COVID-19 Spread (COVIDAi): Project Review

Anđela Blagojević, Tijana Šušteršič, Ivan Lorencin, Nenad Filipović

In this paper, a review of the project Use of Regressive Artificial Intelligence and Machine Learning Methods in Modelling of COVID-19 Spread (COVIDAi) is presented. The main goal of the project is to design two main AI-based models: epidemiological and personalized. After the introduction, a brief description of project partners and activities is provided. Furthermore, a brief description of the two main project activities is provided. After the description of the aforementioned project activities, a review of scientific papers published during project execution is presented.

Investigation of the association between polymorphisms in the circadian CLOCK and NPAS2 genes and cancer by using methods of AI

Ana Mioč, Barbara Fabulić, Jelena Musulin, Daniel Štifanić, Elitza Markova-Car

Circadian rhythm is a natural process in all living organisms that regulates the sleep–wake cycle and repeats roughly every 24 hours. On molecular level, that process is controlled by so-called clock genes. Disruption of circadian rhythms or expression of clock genes is emerging as a novel and potentially modifiable cancer risk factor. Single nucleotide polymorphism (SNP) stands for single base change in a DNA sequence, with an usual alternative of two possible nucleotides at a given position. In this study, the data from case-control studies containing available genotype frequencies of the SNPs in two clock genes (CLOCK and NPAS2) were collected. Based on that data, the association between genetic variations in clock genes and the risk of developing cancer was investigated. Furthermore, Artificial Intelligence (AI) algorithm was developed to predict the type of cancer (breast or mixed).

Multiclass Classification of Oral Squamous Cell Carcinoma

Jelena Musulin, Daniel Štifanić, Ana Zulijani and Zlatan Car

Oral cancer (OC) is a type of head and neck cancer in which malignant cells appear on the lips or in the oral cavity (in the mouth). Early detection of OC may increase the chances of survival in individuals, but new technologies may be expensive and time-consuming. Lately, the possibility of automated medical diagnosis with the aid of Artificial Intelligence (AI) tools has been receiving much attention. In this research, the integration of preprocessing techniques along with Xception algorithm is proposed for oral squamous cell carcinoma classification. The dataset was obtained from the Clinical Hospital Center in Rijeka and consists of 257 histopathology images. The proposed system achieved satisfactory results in terms of multiclass classification.

Z4 HPC Cluster

Sandi Baressi Šegota, Nikola Anđelić, Ivan Lorencin, Daniel Štifanić, Jelena Musulin, Zlatan Car

The computational complexity of research tasks is ever growing, which is something that is extremely apparent in the field of Artificial Intelligence. These computational tasks require High Performance Computers (HPC), which may either be rented, per Infrastructure as a Service (IaaS) paradigm, or purchased in entirety and installed locally. One local cluster is Z4 HPC Cluster installed at the Department of Automation and Electronics, Faculty of Engineering – University of Rijeka. The overview of the hardware and software of the mentioned cluster is given in this paper.