A student-built deep learning model for chest X-ray screening.

A final-year college project exploring on-device medical imaging. Pulmo classifies a chest X-ray into COVID, Normal, Pneumonia, or Tuberculosis , built for learning and demonstration, not clinical use.

/ Analyzer

Drop a chest scan.
Get a verdict in milliseconds.

Idle

Drop an X-ray here

or click to browse · JPG/PNG · processed locally in your browser

Diagnosis

Upload a chest X-ray to see class probabilities.

Research preview only. Not a medical device. Predictions must be reviewed by a qualified clinician before any clinical decision.

/ How it works

A real model. Zero infrastructure.

On-device inference

EfficientNet-B2 backbone with a CBAM attention head, converted to TensorFlow.js graph format and shipped as static shards.

Private by construction

Your scan is decoded into a tensor in your browser. No upload, no server round-trip, no logs.

Calibrated confidence

Softmax probabilities across all four classes, not just a hard label, so you can see when the model is uncertain.