Siamese network ensemble for visual tracking
MetadataShow full item record
Visual object tracking is a challenging task considering illumination variation, occlusion, rotation, deformation and other problems. In this paper, we extend a Siamese INstance search Tracker (SINT) with model updating mechanism to improve its tracking robustness. SINT uses convolutional neural network (CNN) features, and compares the new frame features with the target features in the first frame. The candidate region with the highest similarity score is considered as the tracking result. However, SINT is not robust against large target variation because the matching model is not updated during the whole tracking process. To combat this defect, we propose an Ensemble Siamese Tracker (EST), where the final similarity score is also affected by the similarity with tracking results in recent frames instead of solely considering the first frame. Tracking results in recent frames are used to adjust the model for continuous target change. Meanwhile, we combine large displacement optical flow method with EST to further improve the performance (called EST+). We test the proposed EST and EST+ on a standard tracking benchmark OTB. It turns out the average overlap ratio of EST and EST+ increase 2.72% and 3.55% respectively compared with SINT on OTB 2013, which contains 51 video sequences. For the OTB 100, the average overlap ratio gain is 4.2%.
Showing items related by title, author, creator and subject.
Veronesi, D (2014)The paper examines the phenomenon of correction of musical action by analysing ensemble music workshops devoted to Conduction®, a way of making music together on the basis of a codified gestural lexicon addressed by a ...
Ravazzolo, F; Vahey, SP (2014)We extend the "bottom up" approach for forecasting economic aggregates with disaggregates to probability forecasting. Our methodology utilises a linear opinion pool to combine the forecast densities from many disaggregate ...
Ravazzolo F; Vahey SP (Reserve Bank of Australia, 2010)We construct ensemble predictives for inflation in Australia based on the out-of-sample forecast performance of many component models, where each component model uses a particular disaggregate inflation series. Following ...