World Wide Web of Plankton Image Curation (www.pic)

• Grant number: 18/24167-5
• Support Opportunities: Regular Research Grants
• Start date: June 01, 2019
• End date: November 30, 2023
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Agreement: Belmont Forum
• Principal Investigator: Nina Sumiko Tomita Hirata
• Grantee: Nina Sumiko Tomita Hirata
• Principal researcher abroad: Jean-Olivier Irisson
• Institution abroad: Université Paris-Sorbonne (Paris 4), France
• Host Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated researchers: Roberto Hirata Junior ; Rubens Mendes Lopes
• Associated scholarship(s): 23/10770-0 - Testing and validation of interactive methods for plankton image lebeling, BP.TT; 23/00205-3 - Technical support to the development of efficient methods for plankton image classification, BP.TT; 22/15744-4 - Testing and validation of interactive methods for plankton image lebeling, BP.TT; + associated scholarships 22/09376-2 - Data efficient methods for plankton image classification, BP.PD; 21/02902-8 - Computational technical support for the project World Wide Web of Plankton Image Curation (WWW.PIC), BP.TT; 20/15170-2 - Data efficient methods for plankton image classification, BP.PD - associated scholarships

Abstract

Scientific research is generating an increasing number of digital images, from micrographs of cells to pictures of galaxies. Automated instruments can capture many images, which are then processed automatically to extract data from them. For this data to be useful to the scientific community and benefit the general public, it needs to be fast to generate (even for millions of images), consistent, and easy to share. In environmental sciences, common questions are: how many organisms are present in a given environment? How diverse are they? Does that change in time? Digital imaging can help answer these questions, particularly underwater, where direct observation by humans is difficult. For example, images of billions of planktonic organisms (i.e., the organisms that drift with ocean currents) have been taken and need to be analysed. Plankton largely contributes to the regulation Earth's climate, the production of the oxygen we breathe, the feeding of the fish we eat, etc. Estimating its abundance and diversity is therefore critical. Yet, efforts to process and classify images for such ecological studies have been scattered and not interoperable. The main goal of this project is to build a World Wide Web of Plankton Image Curation applications (WWW.PIC) that collect images of plankton, allow scientists to name them consistently, store associated ecological information (such as time, location, etc.), and make all data easily accessible to the community. It will leverage cutting edge advances in database design and machine learning to process billions of images, will be hosted on public web servers to be easily accessible, and will foster an atmosphere of collaboration and sharing that the Belmont Forum values and, we think, is essential for the progress of science. Then, we will use this network of applications to tackle studies that have proved challenging without it, such as fast plankton monitoring to assess ecosystem health, or global estimations of the distribution of planktonic diversity and its contribution to carbon storage or ecosystem productivity. (AU)