Multi-cancer detection from medical scans in seconds.
CurieSense AI analyzes radiology and pathology images with state-of-the-art neural networks to flag abnormal or cancerous regions, supporting clinicians with fast, consistent second opinions.
Built for accuracy and clinical trust
A complete imaging intelligence layer that turns raw scans into actionable, explainable insights.
Multi-cancer coverage
A single platform trained across multiple modalities — mammography, CT scans, and dermoscopy.
Deep learning models
Convolutional and transformer-based architectures trained on large, annotated medical imaging datasets.
Region localization
Heatmaps and bounding boxes highlight exactly where the model detects abnormal or suspicious tissue.
Fast inference
Results in seconds, enabling rapid triage and a consistent second read for high-volume workloads.
Confidence scoring
Every prediction is returned with a calibrated probability so clinicians can prioritize review.
Privacy-aware
Designed with secure handling of sensitive medical images and clear research-use boundaries.
From scan to insight in three steps
A streamlined workflow designed to fit into clinical and research routines without friction.
Upload a scan
Drag and drop a medical image — mammogram, CT slice, or pathology tile.
AI Analysis
The deep learning model processes the image and localizes abnormal or cancerous regions.
Review results
Get a clear prediction, confidence score, and visual overlay to support your decision.
Measurable performance you can build on
Figures reflect internal benchmark evaluations and are intended for research and decision-support contexts.
- Cancer types & modalities supported
- 3+Cancer types & modalities supported
- Average inference time per scan
- <10sAverage inference time per scan
- Validation sensitivity on benchmark sets
- 94%Validation sensitivity on benchmark sets
- Consistent, fatigue-free second reads
- 24/7Consistent, fatigue-free second reads
Meet the people behind CurieSense AI
University of Kelaniya · Faculty of Computing and Technology · 2025 – 2026
Frequently asked questions
What kinds of images can the system analyze?
The platform is designed to work across multiple modalities including mammograms, chest CT images, dermoscopy (skin) scans, oral photographs, and histopathology slides.
Is this a replacement for a doctor?
No. CurieSense AI is a decision-support and research tool. Its outputs are intended to assist qualified clinicians and should never be used as a sole basis for diagnosis or treatment.
How does the model explain its predictions?
Alongside a class prediction and confidence score, the system produces visual overlays — heatmaps and bounding boxes — that show which regions influenced the result.
Is the project production-ready?
The system is under active development. Core detection capabilities are being refined and expanded across additional cancer types and datasets.
Ready to analyze a scan?
Upload an image and get an AI-assisted multi-cancer assessment in seconds.