eduzhai > Applied Sciences > Engineering >

COMET Context-Aware IoU-Guided Network for Small Object Tracking

  • Save

... pages left unread,continue reading

Document pages: 17 pages

Abstract: We consider the problem of tracking an unknown small target from aerialvideos of medium to high altitudes. This is a challenging problem, which iseven more pronounced in unavoidable scenarios of drastic camera motion and highdensity. To address this problem, we introduce a context-aware IoU-guidedtracker (COMET) that exploits a multitask two-stream network and an offlinereference proposal generation strategy. The proposed network fully exploitstarget-related information by multi-scale feature learning and attentionmodules. The proposed strategy introduces an efficient sampling strategy togeneralize the network on the target and its parts without imposing extracomputational complexity during online tracking. These strategies contributeconsiderably in handling significant occlusions and viewpoint changes.Empirically, COMET outperforms the state-of-the-arts in a range of aerial viewdatasets that focusing on tracking small objects. Specifically, COMEToutperforms the celebrated ATOM tracker by an average margin of 6.2 (and 7 )in precision (and success) score on challenging benchmarks of UAVDT,VisDrone-2019, and Small-90.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...