Adam Blokus, Adam Brzeski, Jan Cychnerski, Tomasz Dziubich, Mateusz Jędrzejewski
https://link.springer.com/chapter/10.1007/978-3-642-30662-4_4
The article presents a novel approach to medical video data analysis and recognition. Emphasis has been put on adapting existing algorithms detecting lesions and bleedings for real time usage in a medical doctor’s office during an endoscopic examination. A system for diagnosis recommendation and disease detection has been designed taking into account the limited mobility of the endoscope and the doctor’s requirements. The main goal of the performed research was to establish the possibility of assuring the necessary performance to introduce the solution into real-life diagnostics.
The structure of an exemplary algorithm has been analyzed to distinguish and discuss parallelization options. After introducing them to the algorithm, the usage of a supercomputer multimedia processing platform allowed to acquire the throughput and latency values required for real-time usage. Four different configurations of the algorithm have been tested and all their measured parameters have been provided and discussed.
