Ph.D. started in: 2018
Expected year of graduation: 2021
COINS consortium member: University of Tromsø
Supervised by: Håvard D. Johansen, Pål Halvorsen, Michael Riegler, Dag Johansen
Research area: Privacy
Project title: Secure and Privacy Preserving Machine-Learning Methods for Disease detection and reporting on Medical Multimedia Data
Project description: Big healthcare data has considerable potential to enhance patient’s results, forecast outbreaks of epidemics, decrease the cost of healthcare delivery and enhance the standard of life in general. Recent decades have seen the emergence of continuous threats, targeted strikes against the information systems. The hospitals record, store and process large amount of medical data but the medical experts are not seen benefiting from it. Although, there has been a lot of promising progresses in this field of pathological disease detection in gastrointestinal tract (GI) tract but it fails to provide a good performance on a real-time system and also fails to solve the problem of security and privacy of patient’s data. Machine learning, big data, and artificial intelligence (AI) are currently three most applicable and trending bit for innovation and predictive analysis in healthcare. The newest generation of machine-learning algorithms (deep learning) may use large standardized data sets produced in healthcare to enhance the efficacy of public health interventions.
The main objective of the project is to develop a medical multimedia system that integrates and combines state-of-the-art tools with new and enhanced algorithms for detection and localization of pathological endoscopic findings and anatomical landmarks in the GI tract. Additionally, we are also focused on finding the new strategies to fix the issues of medical healthcare including challenges in security and privacy.
- Debesh Jha, Vajira Thambawita, Konstantin Pogorelov, Michael Riegler, Pål Halvorsen, Håvard D. Johansen, Dag Johansen (2018). Automatic Disease Detection in Videos Recorded in the Gastrointestinal Tract