VISUAL COMPUTATION: Images and Visualization

Computer vision and digital imaging have a fundamental role in the research envisaged for the USP-e-Science Network: i) digital images retain information covering an array of phenomena in various disciplines; ii) data visualization is one of the most influential tools promoting multidisciplinary interaction for data interpretation by team specialists in a range of scientific domains. Visualization may provide a common language for interdisciplinary groups analyzing case studies, interpreting results and seeking conclusions. The USP-e-Science team members have extensive experience and research projects dealing with these topics.


Pattern recognition methods aim at classifying elements of interest based on experimental measures extracted from them. It plays a central role in most data analysis problems involved in the USP-e-Science center. Pattern recognition applications may be found in a many different fields such as in image analysis, signal processing and computational biology. Among the modern challenges faced by the researchers in the field, it is worth mentioning 3 research topics of particular interest for the USP-e-Science: i) problems where the different elements should be structurally described as networks with connections linking those elements; ii) dynamic problems where the information may evolve along some independent variable (e.g. time in context of the video sequences); iii) problems that present both structural and dynamic aspects, i.e. a network of elements presenting a dynamic evolution along some independent variable.


Due to the nature of this proposal, the Data Mining and Databases group will be one of the structured groups of the USP-e-Science Research Network. Members of this NAP have participated in multidisciplinary projects, which involve the creation and analysis of databases, involving the mining of large databases. In collaboration with Fundação Pró-Sangue Hemocentro de São Paulo they are pursuing the knowledge to improve the security in blood transfusions. The project aims at developing a database to help blood banks (hemocenters) to support analysis of their stored data to gather the information and knowledge needed by the specialists. Three sub-projects will be developed in this direction: i) analysis of residual risk of HIV transmission; ii) epidemiological study regarding the confidence of questionnaires answers obtained during blood donors selection; iii) natural history of the Chagas disease. It is also under development another project in collaboration with the Health Department of Brazil aimed at developing methods and tools to create and to analyze clinical and molecular data markers from HIV patients to help on decision making for the Department of STD (sexual transmitted diseases), Aids and Viral hepatitis.


Research projects related to biological and medical sciences involve modeling and extracting knowledge from biological data of different scales in an integrative approach from systems biology. The involved groups have extensive experience and technology for data acquisition in different biological scales, from genetic to phenotypic levels (especially through images related to the theme of bioimage informatic).