State-Aware Anti-Drift Object Tracking

Abstract

Correlation filter (CF) based trackers have aroused increasing attentions in visual tracking field due to the superior performance on several datasets while maintaining high running speed. For each frame, an ideal filter is trained in order to discriminate the target from its surrounding background. Considering that the target always undergoes external and internal attributes during tracking procedure, the trained filter should take consideration of not only the external distractions but also the target appearance variation synchronously. To this end, we present a State-aware Anti-drift CF tracker (SAT) in this paper, which joint model the discrimination and reliability information in filter learning. Specifically, global context patches are incorporated into filter training stage to better distinguish the target from backgrounds. Meanwhile, a color-based reliable mask is learned to encourage the filter to focus on more reliable regions suitable for tracking. We show that the proposed optimization problem could be solved using Alternative Direction Method of Multipliers and fully carried out in frequency domain to speed up. Extensive experiments are conducted on OTB-100 datasets to compare the SAT tracker (both hand-crafted feature and CNN feature) with other relevant state-of-the-art methods. Both quantitative and qualitative evaluations further demonstrate the effectiveness and robustness of the proposed work.

Aggregation Signature for Small Object Tracking

“…The small90 benchmark: In Fig. 7, we further show the precision and success plots of 30 state-of-the-art trackers including SiamRPN [29] [30], LDES [31], SAT [32], TLD [3], LCT [33], OCT [22], CSK [34], CT [35], STC [36], KCF [37], ECO [38], MDNet [39], LCCF [40], SRDCF [41] and CPF [42], generated by the benchmark toolbox. While several baseline algorithms, e.g., LDES, DaSiamRPN, ECO, have shown promising potential in tracking small objects, our AST still helps achieve the precision rates of 84.9% (LDES AST), 83.1% (DaSiamRPN AST), 83.2% (ECO AST) which improve its counterpart base trackers by 1.6%, 0.9%, 1.7% respectively.…”

Parallel Three-Branch Correlation Filters for Complex Marine Environmental Object Tracking Based on a Confidence Mechanism

“…At this time, the highest peak is not the actual object position. As shown in Figure 1 c, taking the highest peak position as the object position will cause tracking failure such as object drift [ 27 ]. Therefore, the confidence metric proposed in this paper judges whether to update the filter, that is, to supervise the update of the sample model.…”

Towards long‐term UAV object tracking via effective feature matching

“…(ii) Tracking condition judging : Since tracking algorithm often suffers from model drift when occlusion occurs, it is necessary to monitor the tracking status and determine whether to start re‐detection module. Some existing works employ response score or its variant [6, 7 ] as the tracking condition indicator. We argue that the interpretability between occlusion and response value is weak, since the occluder may also provide high response score.…”

“… is the pre‐defined threshold to separate the successful matching from the failed.We select the features which appear steadily as the member in the local feature library K and delete the transient features whose confidence is less than the threshold to maintain a reasonable capacity for the library. (ii) Tracking condition judging : Since tracking algorithm often suffers from model drift when occlusion occurs, it is necessary to monitor the tracking status and determine whether to start re‐detection module. Some existing works employ response score or its variant [6, 7 ] as the tracking condition indicator. We argue that the interpretability between occlusion and response value is weak, since the occluder may also provide high response score.…”