NEO-Healthy: Identification of hypoxic pregnancy conditions in-utero.
A novel risk-scoring tool to identify pregnancies with a high risk of fetal and newborn hypoxia. The process of risk estimation relies on data from electronic health records, laboratory tests, maternal biometric data, claims, patient-generated health data, and the social determinants of health. We use the predictive modelling technique to identify proactively patients with the highest risk of fetal pathology.
The MeCo: A multimodal smartphone-based system for long-term monitoring of patients with Parkinson’s disease.
The MeCo is a 4-year research study that provides freeware for audio, video, and textual analytics to guide patients and their carers to conduct the tests at home and in their homes. The installation package can be downloaded here
CIN-Eva: The AI package to identify cervical intraepithelial neoplasia and carcinoma in situ.
Deep Learning-based automated evaluation tool for cancer screening from cervical images.
Acne-T: A deep oversampling technique for 4-level acne classification that enables to deal with imbalanced datasets.
The proposed approach includes image pre-processing, data augmentation, oversampling, feature extraction, training and model evaluation.
The AIR: A platform that enables researchers to assess their decision-making solutions and AI applications.
The AIR supports clinical development, testing and validation of AI-based solutions. It integrates the most powerful features and outputs of ML platforms to help build the next-generation artificial intelligence in healthcare.
The research aimed to improving the data analysis techniques to find dependencies between the individual factors isolated pregnancy and the presence of the disease in the newborn, discovering interesting associations in the gestation course data, as well as obtaining new grouping factors for the data.
NEO-Healthy is a novel risk-scoring tool to identify pregnant women with a high risk of fetal and newborn hypoxia.
The first implementation of NEO-Healthy was NEOMed (beta release, 2017). After two years of data gathering, digitalization and pre-processing we applied a series of data mining techniques to identify meaningful factors of hypoxic fetal injuries and developed a risk score tool for fetal hypoxia. The process of risk estimation relies on data from electronic health records, lab testing, biometric data, claims, patient-generated health data, and the social determinants of health. We use the predictive modelling technique to identify proactively patients with the highest risk of fetal pathology.
The NEO-Healthy is a further enhancement of NEOMed. We use non-invasive fetal heart monitoring (fHR) to detect abnormal fHR patterns that may indicate oxygen deficiency or other pathological processes at their early stage to perform the appropriate intervention and prevent fetal death and damage.
Our current goal is to automate the detection of the risk of fetal hypoxia through fECG monitoring derived from the signal recorded from the maternal abdominal surface. To solve this task, we propose a novel method for automatic electrocardiogram (fECG) signal processing and hypoxia risk estimation. To the best of our knowledge, this is the first attempt to introduce auto-generated risk scoring in fECG. We assume that risk evaluation allows achieving early warning on fetus’ non-reassuring fetal stat
- Biloborodova T., Scislo L., Skarga-Bandurova I., Sachenko A., Molgad A., Povoroznjuk O., Yevsieiva Y. (2021) Fetal ECG signal processing and identification of hypoxic pregnancy conditions in-utero [J]. Mathematical Biosciences and Engineering, vol. 18(4): 4919-4942. doi: 10.3934/mbe.2021250
- Skarga-Bandurova I., Biloborodova T., Nesterov M. (2019) Extracting Interesting Rules from Gestation Course Data for Early Diagnosis of Neonatal Hypoxia. Journal of Medical Systems. 43:8 Springer US.
- Skarga-Bandurova I., Biloborodova T., Skarha-Bandurov I., Zagorodna N., Shumova L. (2019) EEG Data Fusion for Improving Accuracy of Binary Classification. ICT for Health Science Research A. Shabo (Shvo) et al. (Eds.) Vol. 258. pp. 130-134.
- Skarga-Bandurova I., Biloborodova T., and Dyachenko Y. (2018) Strategy to Managing Mixed Datasets with Missing Items, In: J. Medinaetal. (eds.) Information Processing and Management of Uncertainty in Knowledge-Based Systems – Theory and Foundations. IPMU 2018, Communications in Computer and Information Science, vol. 854. pp. 608–620, 2018. Springer, Cham. https://doi.org/10.1007/978-3-319-91476-3_50
- Skarga-Bandurova I., Biloborodova T., Nesterov M. (2017) Discovering Interesting Associations in Gestation Course Data, Progress in Artificial Intelligence. EPIA 2017. Oliveira E. et al. (Eds): Lecture Notes in Computer Science. vol 10423. pp. 204-214. Springer, Cham. https://doi.org/10.1007/978-3-319-65340-2_17
- Biloborodova T. and Skarga-Bandurova I. (2017) Approaches for Classification of Imbalanced and Skewed Datasets, Herald of East Ukrainian National University. – Severodonetsk: EUNU. No. 8 (238), pp. 17-24.
- Skarga-Bandurova I. and Biloborodova T. (2016) Exploratory Data Analysis to Identifying Meaningful Factors of Hypoxic Fetal Injuries Herald of the National Technical University “KhPI”. Subject issue: Information Science and Modelling. Kharkiv: NTU “KhPI”.
- Skarga-Bandurova I. and Biloborodova T. (2015) Data analysis for prognostication in cerebral status of newborns Herald of the National Technical University “KhPI”. Subject issue: Information Science and Modelling. No. 33 (1142), pp. 134-143.
- Skarga-Bandurova I., Biloborodova T., Nesterov M. (2015) Modeling Structures for Integrated Obstetrics, Gynecology, and Neonatal Information System, Journal of Problems of Information Technologies. No. 17, pp. 51-57.
Control of motor activities is a key clinical need for patients with Parkinson’s disease (PD). Such symptoms are difficult to assess in the outpatient department, and their assessment is often based on unreliable quizzes and questionnaires.
We developed an affordable and easy-to-use solution for remote monitoring of symptoms of PD at home for a continuous long-term assessment of motor conditions.
The present work is in line with the efforts toward predictive healthcare data analytics in capturing early signs of PD as well as assessing the development of Parkinsonian symptoms over time. The idea of this approach involves a combination of well-known methods for testing patients with PD, namely specially designed tests built-in as an application in a smartphone and enable to perform analysis of data in different time and population scales.
The MeCo system includes the following components:
1) A smartphone app for iOS and Android that enables patients to carry out different motor performance tests complete a self-assessment questionnaire, and securely submit data to the cloud service.
2) A web application to collect data from wearable sensors.
3) Scalable cloud-based data collection services that collect data from patients’ smartphones ensure secure data management and apply the data processing pipeline.
4) A web server to host the site and store the signals.
5) Data-mining package for health data analytics incorporating quantitative and semi-structured data, and longitudinal analyses, clustering and classification, and a clinical user interface including short term and long term data visualization.
The raw data from all sensors are processed and then used to perceive the current state. This process is mainly composed of three steps: collection, processing and analysis. These steps are explained below.
The MeCo mobile application is created to control motor fluctuations, their dependence on time, a mode of the day, a physical and psychological condition, type and intensity of medical treatment.
Smartphone-Based System for Long-Term Monitoring of Patients with Parkinson’s Disease
- Biloborodova T., Skarga-Bandurova I., Skarha-Bandurov I. (2021) Knowledge and Data Acquisition in Mobile System for Monitoring Parkinson’s Disease. Information and Knowledge in Internet of Things, EAI Springer Innovations in Communications and Computing Series – in press.
- Biloborodova T, Skarga-Bandurova I, Kotsiuba I, Skarha-Bandurov I. (2021) Reputation-Aware Data Fusion for Quantifying Hand Tremor Severity Form Interaction with a Smartphone. Stud Health Technol Inform. 2021 May 27;281:839-844. DOI: 10.3233/SHTI210297. PMID: 34042792.
- Biloborodova T., Skarga-Bandurova I., Berezhnyi O., Nesterov M., Skarha-Bandurov I. (2020) Multimodal Smartphone-Based System for Long-Term Monitoring of Patients with Parkinson’s Disease. In: Rocha Á., Ferrás C., Montenegro Marin C., Medina García V. (eds) Information Technology and Systems. ICITS 2020. Advances in Intelligent Systems and Computing, vol. 1137. Springer, Cham. https://doi.org/10.1007/978-3-030-40690-5_60
- Biloborodova T., Skarga-Bandurova I., Derevyanchenko V., Skarha-Bandurov I., Tatarchenko H., Mokhonko V. (2019) A Personal Mobile Sensing System for Motor Symptoms Assessment of Parkinson’s Disease. IEEE 7th International Conference on Future Internet of Things and Cloud Workshops (FiCloudW), 2019. pp. 146-151.
Wearable and Embedded IoT-based applications
- Kotsiuba I., Nesterov M., Yanovich Yu., Skarga-Bandurova I., Biloborodova T., Zhygulin V. (2018). Multi–Database Monitoring Tool for the E–Health Services. 2018 IEEE International Conference on Big Data, December 10-13, 2018, Seattle, WA, USA, pp. 2442-2448.
- Skarga-Bandurova I., Biloborodova T. (2019) Wearable and Embedded IoT-based Solutions for Biomedical Applications in Internet of Things for Industry and Human Applications.Volume 3. Assessment and Implementation. IoT for Healthcare systems. Section 46. (Ed. V. S. Kharchenko). Ministry of Education and Science of Ukraine, National Aerospace University KhAI, pp. 535-575.
- Skarga-Bandurova I., Biloborodova T. (2019) Devices with Reconfigurable Architecture for Biomedical IoT-based Applications in Internet of Things for Industry and Human Applications. Volume 3. Assessment and Implementation. IoT for Healthcare systems. Section 47. (Ed. V. S. Kharchenko). Ministry of Education and Science of Ukraine, National Aerospace University KhAI, pp. 576-596.