PROJECT TITLE :
GReTA-A Novel Global and Recursive Tracking Algorithm in Three Dimensions
Tracking multiple moving targets permits quantitative measure of the dynamic behavior in systems as various as animal teams in biology, turbulence in fluid dynamics and crowd and traffic management. In three dimensions, tracking several targets becomes increasingly hard since optical occlusions are very likely, i.e., two featureless targets frequently overlap for many frames. Occlusions are notably frequent in biological teams like bird flocks, fish colleges, and bug swarms, a reality that has severely restricted collective animal behavior field studies within the past. This paper presents a 3D tracking methodology that is strong in the case of severe occlusions. To make sure robustness, we have a tendency to adopt a international optimization approach that works on all objects and frames directly. To realize practicality and scalability, we employ a divide and conquer formulation, due to which the computational complexity of the matter is reduced by orders of magnitude. We tend to tested our algorithm with artificial data, with experimental information of bird flocks and bug swarms and with public benchmark datasets, and show that our system yields prime quality trajectories for tons of moving targets with severe overlap. The results obtained on very heterogeneous data show the potential applicability of our method to the most diverse experimental situations.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here