Stomata

Stomata are microscopic features (comparable to the diameter of a human hair) present on the leaf surfaces of most land plants consisting of two cells surrounding a pore. When enough water is available to the plant the pores will open allowing water vapor to be transpired into the air and carbon dioxide from the air to be imported into the plant. Both the size and number of stomata determine the amount of water transpired and both these traits can vary in time and space in response to climate.

Tropical African rain forest

Although tropical rain forests cover only about 16 percent of the land surface they comprise 50 percent of the terrestrial biodiversity on Earth and store about 250 billion tons of carbon which makes them a key players in mediating climate change. Plants are at the base of the food chain just because for acquiring biomass they only need air (carbon dioxide) and water. Hence, the most fundamental aspects of their biology going from the amount of litter and flowers they produce to the volume of water they evaporate influences the entire tropical rain forest ecosystem and even the entire global carbon-water cycle.

Scientific data from tropical Africa is scarce. The African rain forest is the second largest on Earth yet less than five percent of the total amount of ecological data known for tropical forests are available for tropical Africa.

Tropical forests in general are very inaccessible and still hold many secrets to science. Imagine standing in the middle of a tropical forest surrounded by forest giants and you need to sample leaves from a Prioria species with a crown at 30 meters... .

Fortunately, many of such leaf samples are kept in herbarium collections from museums and botanic gardens worldwide, stored in exactly the right conditions for preserving the herbarium specimens for centuries.

The general aim of the "Leaf Prints" project is to open up a century of ecological data locked up in a herbarium collection from the Meise Botanic Garden in Belgium containing dried plant material collected during the Belgian occupation of Congo (1908-1960). To investigate the adaptation of leaves of tropical tree species to a changing climate we want to count and measure the size and number of leaf stomata and compare these measurements between plant material collected a century ago and very recent material.

Leaf prints can be easily prepared using transparent nail polish and contain plenty of biological and ecological information in which environment the plant lived and even about the plant's position in the flowering plant family tree, all by looking at the shape, size and density of the stomata. For example, stomata from species belonging to the monocots are organized in rows versus scattered in dicots and are usually bigger than the dicot stomata.

Artificial intelligence for automatic measurements of stomata

Counting and measuring of these stomata, however, is very time-consuming. Current solutions using machine learning allow for the automatic recognition of image features, such as stomata. However, these methods rely on very large datasets of validated training data. In this project we will rely on you to provide us with an expert assessment of both the location and the characteristics of stomata across a range of different species (with different stomata shapes and densities) in order to automate future stomata classifications. AS such, other researchers can use the newly developed software to perform their own stomata measurements with only limited human input, speeding up science and discovery relying on stomata measurements.