Image Processing and Mathematical Modelling of Plant Leaf Form and Development

This project will contribute to consolidating the link between Kingston University and RBG, Kew and will provide a significant opportunity to simultaneously push forward scientific frontiers in image processing and botanical research. It is expected that the project will be a valuable source of input to the two teams' research communities, with applications that will extend beyond the initial context of this proposed project. Work on a web-based repository of information, data and extracted features as a resource for other scientists has already taken place and this would be further developed and launched as part of the proposed project.Dr Remagnino shares a common interest with Kew biologists (Drs Wilkin, Mayo and Kirkup) in image understanding and pattern recognition as applied to plant organs . Numerous exchanges of visits have uncovered a common belief in methods and mathematical techniques for the automatic interpretation of plant form and development. Kew and its collections comprise 25,000 species of living plant and seven million stored dried specimens, of which over 100,000 have been imaged. It is therefore timely for Dr Remagnino and Drs Wilkin, Mayo and Kirkup to explore at greater depth botanical research and design novel algorithms for image interpretation.The professional scientific relationship between Dr Remagnino and the biologists at the RBG, Kew has matured over the last few years. It is now time for Dr Remagnino to set aside a longer period of time to design image understanding algorithms for the classification of leaves and modelling of their growth and development. The Discipline Hopping award would provide the required time to carry out the study and spend time with the Kew biologists for discussions and implementation of ideas that have been formed over the past few years. The award would give Dr Remagnino the opportunity to study in greater depth the biological aspects of leaf classification, gaining a detailed understanding of the problem. The award would provide an invaluable vehicle to investigate a number of hypotheses and algorithms on how leaf images can be analysed and a thorough analysis of the outcomes.