PIPELINE
How LungAI
processes a scan.
A trained CNN runs convolutional inference over a 224×224 grayscale CT slice and returns a binary classification with a confidence score.
01
Upload
A CT slice is sent to the FastAPI backend as multipart form data.
02
Preprocess
Image is converted to grayscale, resized to 224×224, and normalized to [0, 1].
03
Convolution
A small CNN runs two Conv2D + MaxPool blocks followed by a dense head.
04
Sigmoid
The final layer outputs a probability between 0 and 1.
05
Threshold
Scores above 0.5 are flagged. The raw score is returned for transparency.
06
Response
JSON is returned with label, confidence, raw score, and threshold.
COMPARISON
Why AI-assisted screening matters
Traditional radiology
- Days to weeks for scans to be reviewed
- Subject to human fatigue and oversight
- Limited specialist access in many regions
AI-assisted screening
- Sub-second inference per slice
- Consistent baseline across millions of scans
- Available anywhere there's a browser
See it in action.
Run the model against a CT slice and watch it return a confidence score in real time.
Open analyzer