Domain generalization tutorial

Slides for the domain generalization tutorial at the CV4Africa workshop at the 5th Deep Learning Indaba 2023

Cassava disease classification

Cassava is an important provider of carbohydrates in Africa and is therefore keyto the nutrition of a large part of the population of the continent. The plant can be affected by several diseases that result in lower yields. Identifying them can be beneficial to food security. To help solve this problem, machine learning has been employed to identify diseases in pictures of cassava leaves. The data collected to make this possible was made available in the iCassava-2019 competition 1. During our computer vision class at AMMI we were required to work on this datset as part of an In-Class Kaggle competition which my team (Simon Bassey and I) won. In this blogpost I will describe our thought process and the simple methods we used. Ernest Mwebaze, Timnit Gebru, Andrea Frome, Solomon Nsumba, Jeremy Tusubira, iCassava 2019 Fine-Grained Visual Categorization Challenge (2019), arXiv:1908.02900 ↩