ALUNA (AI Lung Analyzer) is a research-grade web application for detecting and classifying pulmonary nodules in CT scans using a custom-trained YOLOv8 model exported to ONNX format.
Assist radiologists in early-stage lung cancer screening through AI-assisted nodule detection.
Trained on the LIDC-IDRI dataset — the largest public lung CT dataset with expert radiologist annotations.
Next.js 15, ONNX Runtime (Node), Sharp for image processing, and dicom-parser for DICOM support.
For research and educational use only. NOT a clinical decision-support tool. Always consult a radiologist.
| Architecture | YOLOv8 (custom) |
| Export Format | ONNX |
| Input Size | 640 × 640 px |
| Classes | Benign / Equivocal / Malignant |
| DICOM Windowing | WC = −600, WW = 1500 |
| NMS IoU Threshold | 0.45 |
| Default Confidence | 0.25 |