This project proposes the development and application of mathematical and image analysis tools derived from fractal geometry to the analysis and identification of plant species of Brazilian flora. The proposed use of fractal geometry for this analysis comes from the precision that this approach presents in the modeling of objects from nature. Actually, elements such as plant leaves have properties closer to those of a fractal object than to objects of classical geometry. For example, these structures are highly complex and have a significant level of self-similarity in their appearance. This work will explore techniques known in the literature such as multifractal and fractal descriptors and propose new approaches for extracting image features based on fractal geometry. This combination of mathematical and computational tools with biological techniques will provide benefits for both exact and biological areas. In this way, this study allows the development of a tool that can turn the process of identifying and analyzing plants more precise and agile. At the same time, it encourages the development of more advanced techniques in image analysis.
News published in Agência FAPESP Newsletter about the scholarship: