New Approach for Accelerating Disease Detection Using a Parallelism-Nased CNN Model

Abstract

In light of the recent pandemics like COVID phase in 2020 that the world is going through, with all the gaps and incompleteness in diagnostics, as well as the slowness of decision-making for patients, the interest of AI to solve in the first place the various problems linked to diagnosis and secondly having a rapid response to be able to intervene in time to save lives, becomes a priority of priorities.

This work follows work which aimed to reduce the training data necessary in the deep learning process to increase the possibility of its use in cases where time is essential and the contribution was on a basis of extracting features manually with extraction methods and techniques from machine learning combine with deep learning. In our approach, we keep the same logic and we apply or integrate the principle of parallel programming into the process in order to further reduce processing and decision-making time. This will give us a very important advantage in making decisions and helping doctors with diagnosis which will give an advantage over diseases.