Artificial Intelligence Quantitative Imaging Lab
Modern radiology clinics are overwhelmed by growing volumes of DICOM imaging data. Fragmented systems for patient records, image cataloging, and review create bottlenecks that delay diagnoses.
AQUILA unifies image ingestion, study organization, and AI-assisted interpretation in one secure platform — so radiologists can read more studies, faster, with greater confidence.
AQUILA (Artificial Intelligence Quantitative Imaging Lab) is a centralized DICOM imaging workspace built for radiology clinics. It offers secure file intake, automated background processing, and smart study cataloging — all in one place.
An integrated web viewer supports detailed image review and ROI annotation, while a built-in patient registry manages demographics, accession numbers, and FHIR-style terminology. Role-based access control keeps every study secure and collaborative.
AQUILA's Autonomous Imaging Intelligence (AAII) system brings advanced machine learning to clinical reading. MedSigLIP performs zero-shot image classification, scoring studies against radiology candidate labels automatically.
MedGemma generates structured draft radiology reports and powers an interactive Visual Question Answering system — letting clinicians ask free-form questions and track disease progression across longitudinal studies.
Backend: Flask with OpenAPI3 and SQLAlchemy, powered by PostgreSQL for DICOM metadata, AI findings, and FHIR resources.
Frontend: Server-rendered Jinja templates with Vue 3 and Bootstrap 5.
Security: OIDC authentication with role-based access control across 11 clinical roles.
Infrastructure: Asynchronous workers for DICOM parsing and AI inference, with Hugging Face Endpoints, S3-compatible storage, and PACS integration.
Deliver faster, more accurate diagnoses with AQUILA — the secure, scalable AI imaging platform for modern radiology.