Data efficient methods for plankton image classification

• Grant number: 22/09376-2
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: August 01, 2022
• End date: October 31, 2023
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Agreement: Belmont Forum
• Principal Investigator: Nina Sumiko Tomita Hirata
• Grantee: Alexandre Morimitsu
• Host Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated research grant: 18/24167-5 - World Wide Web of Plankton Image Curation (www.pic), AP.R

Abstract

The main objective of this postdoctoral project is to continue the development of machine learning based computational methods to accelerate both the annotation and classification of plankton images, minimizing the required effort from the expert. The method must be able to make effective use of previously generated knowledge, quickly producing new classifiers adapted to new images or usage conditions. Some of the topics to be explored in the project are unsupervised and semi-supervised machine learning approaches, possibly including user-interaction mechanisms, as well as techniques related to novelty detection and class imbalance treatment. (AU)