1 July 2021 in Medical Image Analysis
1 May 2021 in Medical Image Analysis
14 August 2020 in IEEE Transactions on Medical Imaging
Although wireless capsule endoscopy is the preferred modality for diagnosis and assessment of small bowel diseases, the poor camera resolution is a substantial limitation for both subjective and automated diagnostics. Enhanced-resolution endoscopy has shown to improve adenoma detection rate for conventional endoscopy and is likely to do the same for capsule endoscopy. In this work, we propose and quantitatively validate a novel framework to learn a mapping from low-to-high-resolution endoscopic images. We combine conditional adversarial networks with a spatial attention block to improve the resolution by up to factors of 8×, 10×, 12×, respectively. Quantitative and qualitative studies demonstrate the superiority of EndoL2H over state-of-the-art deep super-resolution methods Deep Back-Projection Networks (DBPN), Deep Residual Channel Attention Networks (RCAN) and Super Resolution Generative Adversarial Network (SRGAN). Mean Opinion Score (MOS) tests were performed by 30 gastroenterologists qualitatively assess and confirm the clinical relevance of the approach. EndoL2H is generally applicable to any endoscopic capsule system and has the potential to improve diagnosis and better harness computational approaches for polyp detection and characterization. Our code and trained models are available at https://github.com/CapsuleEndoscope/EndoL2H.
3 April 2020 in IEEE Sensors Journal
High-speed Rotational Angiography scans are prone to potential collisions between the C-arm X-ray system and the patient. A key factor during rotational angiography clinical workflow is thus the initial patient positioning to ensure a collision-free scan. The current practice for this involves multiple manual iterations of a low speed safety-run prior to the actual scan. Several iterations are often required before a scan is collision-free leading to a suboptimal clinical workflow. This work proposes a RGB-D camera based safety system which automates the patient positioning process. The camera determines collisions between the C-arm and patient and a re-positioning algorithm determines the movement of the patient table required to ensure a collisoin-free scan. A unique feature of the solution is the ability to warn clinical staff when a collision-free scan is not possible. The efficacy of the proposed system is illustrated experimentally for two major RA scan types.