Learning from uneven video streams in a multi-camera scenario

We present a semi-supervised incremental learning algorithm for evolving visual data in order to develop a robust and flexible track classification system in a multi camera surveillance scenario. Most existing methods, which are variations on static learning schemes, can not cope with many real-life challenges. The scarcity of labelled data in real applications ends up generating poor classifiers. Furthermore, labelling the whole data (possibly massive in such applications) impose a high cost to the system, rendering the technology impractical. So, there is an increasing interest on exploiting un-labelled data. Our proposed method learns from consecutive batches by updating an ensemble in each time. It tries to achieve a balance between performance of the system and amount of data which needs to be labelled. As no restriction is considered, the system can address many practical problems in an evolving multi-camera scenario, such as concept drift, class evolution and various length of video streams which have not been addressed before.