Informatics Engineering and Computer Science
Recent technological advances made it possible to collect large volumes of data on the evolution of spatio-temporal entities, far superior to the existing capacity to analyze them and extract relevant information. For this reason, tools capable of automating quantitative analysis processes on spatio-temporal data are increasingly needed, guaranteeing levels of objectivity, precision and reproducibility compatible with scientific work. The activities presented next are part of the WP3 of the MOST project and further extend the work previously carried out: Part I - detection of spatio-temporal events To extend the future work presented in a master's thesis on the representation and automatic detection of events related to a spatio-temporal entity, already carried out in this context and project scope. It is intended to: - complement the results already achieved in the aforementioned thesis and expand the knowledge of the state of the art in the area of change detection algorithms, as a key piece in the calculation of differences between pairs of geometric shapes that represent two states of a space-time entity - classify the algorithms and select the most promising algorithm, substantiating the choice - experiment and compare the new results with those previously obtained - write a scientific article presenting the work done and results obtained Part II - develop a test scenario using boids (from computer graphics) - experiment the new test scenario with the already developed prototypes related to spatial and spatio-temporal indexing; - collect and analyze results and draw conclusions - write a scientific article presenting the work done and results achieved Part III - write the scholarship activity report
Master's student in computer science, engineering or similar.
Minimum profile required
- current classification equal or higher than 13 values;- no overdue course units.
Since 25 Nov 2020 to 14 Dec 2020
Cluster / Centre
Computer Science / Information Systems and Computer Graphics