ITS · Machine learning

Showroom of ITS ML cases

Drawing on our machine-learning experience, we expose a public demo of our unified ML pipelines built with Kedro, PyTorch and MLflow.

Pick one of the five cases below to open its demo. To dive deeper into our CI/CD and production-like experimental environment, write to us: presales@its.xyz

Floor Plan Recognition Vectorised floor plans from raster drawings

Walls, windows, doors and fixtures are recognised on a scanned floor plan and redrawn as a clean vector-style scheme.

Discover demo →
Railways Segmentation Track segmentation on high-resolution railway imagery

Two SwiftNet models segment the rails on 2048×1084 frames from a locomotive-mounted camera.

Discover demo →
Polar Bear Detection Polar bears on aerial photography

Classical image filters propose candidate spots on 7360×4912 aerial frames; a MobileNetV3 detector confirms the bears.

Discover demo →
Bacteria Detection Colony detection and classification under the microscope

A Mask R-CNN finds bacterial colonies on microscope images and labels each of the 8 classes with a confidence score.

Discover demo →
Bacteria Segmentation Pixel-level colony masks

A U-Net produces per-pixel masks of bacterial colonies for the same 8 classes as the detection case.

Discover demo →
Eye Microvessels Conjunctival vessel segmentation and measurement

A U-Net segments the microvessels on a slit-lamp photo of the eye; each vessel is then skeletonised, classified as large or medium and measured in micrometres (F-score 0.78).

Discover demo →