GradNet: Gradient-Guided Network for Visual Object Tracking

Abstract

The fully-convolutional siamese network based on template matching has shown great potentials in visual tracking. During testing, the template is fixed with the initial target feature and the performance totally relies on the general matching ability of the siamese network. However, this manner cannot capture the temporal variations of targets or background clutter. In this work, we propose a novel gradient-guided network to exploit the discriminative information in gradients and update the template in the siamese network through feed-forward and backward operations. To be specific, the algorithm can utilize the information from the gradient to update the template in the current frame. In addition, a template generalization training method is proposed to better use gradient information and avoid overfitting. To our knowledge, this work is the first attempt to exploit the information in the gradient for template update in siamese-based trackers. Extensive experiments on recent benchmarks demonstrate that our method achieves better performance than other state-of-the-art trackers. The source codes are available at

LaSOT: A High-Quality Benchmark for Large-Scale Single Object Tracking

“…To alleviate this problem, deep Siamese networks have been introduced for object tracking [6,75]. Owing to balanced efficiency and accuracy, deep Siamese tracking has been extended by many later works [27,36,49,51,78,80,95]. To deal with scale variation, the methods of [15] introduce the intersectionover-union (IoU) network for tracking and achieve promising results.…”

Adaptive Aggregation of Arbitrary Online Trackers with a Regret Bound

“…In relation to experts and comparative methods, twelve state-of-the-art online trackers were employed (ATOM [29], DaSiamRPN [30], GradNet [36], MemTrack [37], SiamDW [38], SiamFC [39], SiamMCF [31], SiamRPN [40], 4. Specifically, this condition means that t + Dt is the same as Z, where Dt is the total delay until t defined as…”

“…In particular, the potential to combine and/or select trackers based on their reliability will make the results more robust than with a single one. ATOM [29] DaSiamRPN [30] GradNet [36] MemTrack [37] SiamDW [38] SiamFC [39] SiamMCF [31] SiamRPN [40] SiamRPN++ [32] SPM [33] Staple [41] THOR [34] Fig. 1: No "almighty" tracker.…”

Complementary Discriminative Correlation Filters Based on Collaborative Representation for Visual Object Tracking

“…State-of-the-art competitors: We compare our CRCDCF with 21 state-of-the-art trackers including SAMF [15], SRDCF [31], Staple [20], BACF [22], STAPLE-CA [21], CSRDCF [52] , ECO-HC [9], STRCF [54], CREST [55], CFWCR [56], SiamFC [57], CFNet [58], MCPF [59], C-COT [60] and ECO [9], SRDCF-deep [17], TRACA [61], MCCT [48], GCT [62], TADT [63], and GradNet [64]. For a fair comparison, all the results of these trackers were obtained by re-running these algorithms on the datasets using the source codes published by the original authors with the provided parameter settings.…”