Convolutional Features for Correlation Filter Based Visual Tracking

Martin Danelljan, Gustav Hager, Fahad Shahbaz Khan, Michael Felsberg; Proceedings of the IEEE International Conference on Computer Vision (ICCV) Workshops, 2015, pp. 58-66

Visual object tracking is a challenging computer vision problem with numerous real-world applications. This pa- per investigates the impact of convolutional features for the visual tracking problem. We propose to use activations from the convolutional layer of a CNN in discriminative correlation filter based tracking frameworks. These acti- vations have several advantages compared to the standard deep features (fully connected layers). Firstly, they miti- gate the need of task specific fine-tuning. Secondly, they contain structural information crucial for the tracking prob- lem. Lastly, these activations have low dimensionality. We perform comprehensive experiments on three benchmark datasets: OTB, ALOV300++ and the recently introduced VOT2015. Surprisingly, different to image classification, our results suggest that activations from the first layer pro- vide superior tracking performance compared to the deeper layers. Our results further show that the convolutional fea- tures provide improved results compared to standard hand- crafted features. Finally, results comparable to state-of-the- art trackers are obtained on all three benchmark datasets.