Geintra

In this paper, the authors propose the use of a probabilistic algorithm to develop a multi-object tracking task. Different solutions have already been proposed by the scientific community to find a solution for this application, and the particle filter is proven to be the best choice in this case as the multimodality character of this Bayes filter implementation can be well-spent: an only particle filter can be used to track a variable number of objects. On the other hand, the flexibility of the particle filter can become in a lack of robustness for the estimator, so a deterministic algorithm is added to the standard filter to increase it. In the following paragraphs the global combined algorithm is described and different tests based on stereo-vision information are also included, proving the reliability of the proposal tracker