e-learning

GLEAM Image Learner - Validating Skin Lesion Classification on HAM10000

Abstract

In this tutorial, we will use the HAM10000 ("Human Against Machine with 10,000 training images") dataset to develop a deep learning classifier for dermoscopic skin lesion classification. The goal is to accurately classify seven types of pigmented skin lesions using the GLEAM Image Learner tool.

About This Material

This is a Hands-on Tutorial from the GTN which is usable either for individual self-study, or as a teaching material in a classroom.

Questions this will address

  • How do we validate GLEAM's Image Learner against a published benchmark on HAM10000?
  • How do we set up a balanced train/validation/test split for multi-class image classification?
  • How do we interpret accuracy, weighted precision/recall, and weighted F1 for imbalanced medical imaging datasets?

Learning Objectives

  • Prepare a balanced HAM10000 subset and perform a stratified 70/10/20 train/validation/test split.
  • Train an Image Learner model using a pretrained CaFormer S18 384 backbone.
  • Evaluate performance using accuracy and weighted precision/recall/F1, and inspect confusion patterns.

Licence: Creative Commons Attribution 4.0 International

Keywords: Deep Learning, HAM10000 Dataset, Image Classification, Image Learner, Skin Lesion Classification, Statistics and machine learning

Target audience: Students

Resource type: e-learning

Version: 1

Status: Active

Learning objectives:

  • Prepare a balanced HAM10000 subset and perform a stratified 70/10/20 train/validation/test split.
  • Train an Image Learner model using a pretrained CaFormer S18 384 backbone.
  • Evaluate performance using accuracy and weighted precision/recall/F1, and inspect confusion patterns.

Date modified: 2026-01-28

Date published: 2026-01-28

Authors: Alyssa Pybus, Jeremy Goecks, Junhao Qiu, Khai Van Dang, Paulo Cilas Morais Lyra Junior

Scientific topics: Statistics and probability


Activity log