eXplainable Artificial intelligence and Virtual reality for Enhanced Radiology (XAVIER)
Type: National Project Project
Duration: from 2023 Jan 01 to 2025 Dec 31
Financed by: FCT
Prime Contractor: R - INESC-ID Lisboa (Other) - Lisboa, Portugal
XAVIER aims to make Deep Learning models understandable to radiologists by promoting Virtual Reality (VR) as a disruptive technology for clinical practice in assessing chest X-ray images. XAVIER will use VR technologies to collect radiologists' eye-tracking and pupil dilations data to design novel multimodal learning architectures that can learn from human classification patterns and generate human-centric explanations. Consequently, XAVIER aligns with the scientific domain of Engineering and Technological Sciences. It has a strong foundation in the engineering advancements of VR headsets and displays resolutions and recent technological and algorithmic developments in deep learning technologies for image classification. XAVIER also aligns with the communication engineering and systems domain. Making deep learning models explainable to humans depends on novel interactive algorithms to generate understandable human explanations to expert radiologists.
Partnerships
- Lusíadas Knowledge Center (Other) - Portugal