2015 - 2017
This project focuses on solving problems related to large-scale visual surveillance, where data is acquired from multiple surveillance cameras. The main hypothesis is that the combination of characteristic descriptors is capable of providing better results for problems within this scope. In addition to supplementing resources for the project already approved by CNPq, financing the proposed work plan will enable the execution of the new proposed activities, which complement the original project. Additionally, it will allow advancement in the area of visual surveillance towards solving high-level problems, such as anomalous movement detection and activity recognition, which are the main focuses of intelligent surveillance systems, where the human operator will receive pre-selected information at the same time. instead of checking multiple monitors for long periods of time as currently, which is susceptible to errors due to human inability to perform repetitive and tedious tasks.