Epidermis thickness

The Epidermis Thickness AI model allows you to estimate the average thickness of the epidermal layer of the skin on H&E stained whole slide images (WSIs).

Training Dataset:

The model was trained on a dataset comprising 114 WSIs containing labeled epidermis layer masks created by professional dermatopathologists.

Performance Metrics:

Intersection over Union (IoU): 0.82
Accuracy: 0.91
DICE: 0.89

Algorithm Description

The algorithm first runs the epidermis segmentation model, then measures and averages the distance between segmented tissue borders. The model returns a segmentation mask with the average epidermis thickness.

Technical Details


  • Compatible with most modern web browsers and mobile devices.

  • Optimal performance is achieved with at least 2GB of RAM and a stable internet connection.