Shayan Sharifi

⌨️ Project Portfolio Archive

Explainable AI for ECG-based Screening of LGE+ Cardiomyopathies

This study aims to develop explainable machine-learning and deep-learning models to identify ECG-derived signatures associated with LGE-positive cardiomyopathies, currently detectable only through contrast-enhanced cardiac magnetic resonance imaging. If successful, the project would provide a low-cost, non-invasive proxy based on routine ECG recordings to support early risk stratification and guide referral to specialised, more expensive and invasive diagnostic pathways only for patients with suspicious ECG profiles. The study is carried out in collaboration with ASUGI–Cattinara.

Multi-modal Explainable 3D Object Detection

We proposed a Trustworthy AI perception module that is remarkably robust, integrates faithful explainability, and calibrated uncertainty estimates. Building on a transformer-based detector, we derive explanation from the attention mechanism at inference time and validate their faithfulness using perturbation-based consistency tests. Experiments show faithful saliency behavior, improved robustness, and well-calibrated uncertainty estimates. Finally, we deploy these Trustworthy AI elements in a prototype vehicle and provide an XAI Interface that visualizes documentation artifacts, model uncertainty state, and saliency maps, demonstrating the feasibility of trustworthy perception monitoring in real time