Automatic initialization and tracking using attentional mechanisms

A biologically inspired approach for automated visual tracking is proposed. In this approach it is hypothesized that target initialization and tracking are a consequence of saliency mechanisms that guide the deployment of visual attention. The recently proposed discriminant center-surround saliency model, is used to derive the tracking framework. In this framework, automatic tracker initialization is achieved using bottom-up saliency with motion features, while the tracking problem is formulated as one of continuous target-background classification, implemented using saliency in two stages. The first, or learning stage, combines a focus of attention mechanism and bottom-up saliency to identify a maximally discriminant set of features for target detection. The second, or detection stage, uses a feature based attention mechanism and a target-tuned top-down discriminant saliency detector, to detect the target. Overall, the tracker iterates between learning discriminant features from the target location in a video frame and detecting the location of the target in the next frame. To implement this tracker, well known properties of the statistics of natural images are exploited leading to computational efficiency. Experimental results comparing the proposed method to the state of the art in tracking are presented, showing improved performance.