ABSTRACT
Catheter tip misalignment can lead to complications in patients together with serious medical malpractice cases. This article aims at the current surge in COVID-19 patients. Using X-ray imaging datasets from COVID-19 patients, previously published on Kaggle as 'RANZCR CLiP - Catheter and Line Position Challenge' and hosted by the Royal Australian and NZ College of Radiologists, a deep-learning algorithm was utilized to detect the position of the patient's catheter and automatically determine whether the catheter tip is misplaced or otherwise. This study employed U-Net to segment and identify catheter position types, together with employing Efficiency net B7 to determine whether the misaligned catheter is misaligned which scores 0.959(AUC). In addition, results were also compared using Efficiency Net B5, ResNet 200D. © 2022 IEEE.