Publications by authors named "Florin-Cristian Ghesu"
Radiofrequency (RF) ablation is a minimally invasive therapy for atrial fibrillation. Conventional RF procedures lack intraoperative monitoring of ablation-induced necrosis, complicating assessment of completeness. While spectroscopic photoacoustic (sPA) imaging shows promise in distinguishing ablated tissue, multi-spectral imaging is challenging in vivo due to low imaging quality caused by motion.
View Article and Find Full Text PDFObjectives: Chest radiographs (CXRs) are commonly performed in emergency units (EUs), but the interpretation requires radiology experience. We developed an artificial intelligence (AI) system (precommercial) that aims to mimic board-certified radiologists' (BCRs') performance and can therefore support non-radiology residents (NRRs) in clinical settings lacking 24/7 radiology coverage. We validated by quantifying the clinical value of our AI system for radiology residents (RRs) and EU-experienced NRRs in a clinically representative EU setting.
View Article and Find Full Text PDFObjectives: Diagnostic accuracy of artificial intelligence (AI) pneumothorax (PTX) detection in chest radiographs (CXR) is limited by the noisy annotation quality of public training data and confounding thoracic tubes (TT). We hypothesize that in-image annotations of the dehiscent visceral pleura for algorithm training boosts algorithm's performance and suppresses confounders.
Methods: Our single-center evaluation cohort of 3062 supine CXRs includes 760 PTX-positive cases with radiological annotations of PTX size and inserted TTs.
Multi-Scale Deep Reinforcement Learning for Real-Time 3D-Landmark Detection in CT Scans.
IEEE Trans Pattern Anal Mach Intell
January 2019
Article Synopsis
- Current methods for detecting anatomical structures in medical images often rely on traditional machine learning techniques that have limitations in feature engineering and search algorithms.
- The proposed method utilizes deep reinforcement learning to create an artificial agent that learns to both identify and locate anatomical objects effectively and efficiently.
- Evaluation of this new approach on substantial medical imaging data shows it significantly improves detection accuracy and speed compared to existing methods, achieving real-time performance with no clinical failures.