Tracking in infrared imagery based on a point of interest/region cooperative segmentation

Describes the results of a work in the field of dynamic scene analysis and more specifically dealing with target tracking from infrared imagery. In order to develop an algorithm which can be considered for the design of future real systems, the authors have tried to satisfy the constraints of time, robustness, accuracy and size in an algorithmical way. This means they have tried to reduce the complexity of a generic envisioned method, and to use the same operators in multiple steps of the algorithm. So they have developed a new segmentation method, deduced from a physical model related to the infrared image formation, and adapted it to the application (ie the target tracking). Another advantage of the algorithm is its modularity: some parts can be used in other systems