| Grant number: | 13/26647-0 |
| Support Opportunities: | Scholarships abroad - Research Internship - Doctorate |
| Start date: | May 01, 2014 |
| End date: | March 31, 2015 |
| Field of knowledge: | Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques |
| Principal Investigator: | Maria Cristina Ferreira de Oliveira |
| Grantee: | Vinícius Ruela Pereira Borges |
| Supervisor: | Bernd Hamann |
| Host Institution: | Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil |
| Institution abroad: | University of California, Davis (UC Davis), United States |
| Associated to the scholarship: | 12/00269-7 - Visual exploration of feature spaces to support Green Algae Taxonomic Classification, BP.DR |
Abstract The taxonomic classification of freshwater green microalgae is highly problematic, with many known inconsistencies. Biologists-taxonomists manually inspect and analyze microscopy images, and typically carry out several complex and meticulous procedures that are time consuming and demand considerable expert knowledge. The problem can greatly benefit from computational tools, in particular the application of techniques in digital image analysis and visual data mining.In this doctoral project we are working with biologists from the Federal University of São Carlos (UFSCar) to handle this problem. Our goal is to define an interactive visual exploration process to support the tasks involved in the classification of green algae at the level of species. This process shall rely on user-assisted classification of images with the support of visualizations based on phylogenetic trees (known as Neighbor-Joining - NJ -trees, and its variations). The NJ-tree representations of data offer a visual metaphor capable of conveying a hierarchical representation of the similarity relations between high-dimensional data. Images from a collection may be represented as feature vectors, from which (dis)similarities may be estimated with appropriate distance functions. Thus, computational challenges for this project include the identification and extraction of relevant features and definition of distance functions suitable to compute their similarity, the integration with a user-driven visual classification system, and the validation of the results obtained with the proposed strategy as compared with current procedures, which are not well-formalized and require considerable implicit knowledge. (AU) | |
| News published in Agência FAPESP Newsletter about the scholarship: | |
| More itemsLess items | |
| TITULO | |
| Articles published in other media outlets ( ): | |
| More itemsLess items | |
| VEICULO: TITULO (DATA) | |
| VEICULO: TITULO (DATA) | |