v1.0.0
ALUNA leverages state-of-the-art deep learning to detect and classify lung nodules from CT scans with high accuracy supporting DICOM, PNG, JPG and BMP formats.
State-of-the-art object detection model trained on the LIDC-IDRI dataset for precise nodule localisation.
Upload raw DICOM files (.dcm) directly with automatic HU windowing (WC=−600, WW=1500) for optimal lung visualisation.
ONNX Runtime on the server side with model singleton caching ensures blazing-fast detection even on CPU.
Every detected nodule is classified as Benign, Equivocal, or Malignant with confidence scores for transparent reporting.
Analyse multiple scans simultaneously. Progress tracking and per-scan status so you never lose sight of the queue.
Bounding boxes are drawn directly onto the CT image with colour-coded labels, ready for download or inspection.
Drag & drop your CT scan (DICOM, PNG, JPG or BMP).
Hit "Run Detection" and the model infers in seconds.
Inspect annotated results with bounding boxes and confidence scores.
Non-cancerous nodules. Typically calcified or with smooth margins.
Indeterminate nodules requiring follow-up or additional imaging.
Suspicious nodules with features consistent with lung cancer.