Despite the numerous developments in object tracking, further development of current tracking algorithms is limited by small and mostly saturated datasets. As a matter of fact, data-hungry trackers based on deep-learning currently rely on object detection datasets due to the scarcity of dedicated large-scale tracking datasets. In this work, we present TrackingNet, the first large-scale dataset and benchmark for object tracking in the wild. We provide more than 30K videos with more than 14 million dense bounding box annotations. Our dataset covers a wide selection of object classes in broad and diverse context. By releasing such a large-scale dataset, we expect deep trackers to further improve and generalize. In addition, we introduce a new benchmark composed of 500 novel videos, modeled with a distribution similar to our training dataset. By sequestering the annotation of the test set and providing an online evaluation server, we provide a fair benchmark for future development of object trackers. Deep trackers fine-tuned on a fraction of our dataset improve their performance by up to 1.6% on OTB100 and up to 1.7% on TrackingNet Test. We provide an extensive benchmark on TrackingNet by evaluating more than 20 trackers. Our results suggest that object tracking in the wild is far from being solved.

Abstract• Correlation filter (CF) trackers use circularly shifted patches as a proxy to real translations when training the filter.• In all CF trackers, correlation scores are regressed to a Gaussian pulse centered around the previous location.• This limits the trackers' ability to recover from partial occlusion or to track fast moving objects.• We propose a generic framework, in which we solve for both the optimal filter and target response based on exact translation detection scores from the image.• The final formulation can use kernels and multiple templates jointly if solved in the dual domain. It improves all baseline CF trackers (SAMF, KCF, DCF, CSK, and MOSSE) by (3 -13)% on OTB100 [1].

No abstract

No abstract

Randomized smoothing is a recent technique that achieves state-of-art performance in training certifiably robust deep neural networks. While the smoothing family of distributions is often connected to the choice of the norm used for certification, the parameters of the distributions are always set as global hyper parameters independent of the input data on which a network is certified. In this work, we revisit Gaussian randomized smoothing where we show that the variance of the Gaussian distribution can be optimized at each input so as to maximize the certification radius for the construction of the smoothed classifier. This new approach is generic, parameter-free, and easy to implement. In fact, we show that our data dependent framework can be seamlessly incorporated into 3 randomized smoothing approaches, leading to consistent improved certified accuracy. When this framework is used in the training routine of these approaches followed by a data dependent certification, we get 9% and 6% improvement over the certified accuracy of the strongest baseline for a radius of 0.5 on CIFAR10 and ImageNet, respectively. Our implementation can be found at https://github.com/MotasemAlfarra/Data-Dependent-Randomized-Smoothing