硕士投稿ECCV2018 论文介绍-Response Map Quality-Aware 高效解决遮挡和目标变形的目标跟踪算法。

              “Response Map Quality-Aware Correlation Filter Tracking”(submitted to ECCV workshop 2018) 完成。论文基于 QMF 和 Log-polar 高效解决目标跟踪算法中的遮挡和尺度变化问题，详见其原始 论文：https://arxiv.org/abs/1807.04880。 该论文在 AAAI2019 under review 状态。
➢ 定义高效 response-map 质量评测函数 QMF,论文从理论上证明其高效性。
➢ 结合 QMF 和特殊模型更新模型，在其他常见 CF trackers 亦取得性能提升（ BACF， Staple, KCF）。
➢ 算法在 VOT2017 real-time sub-challenge 取得 State-of-the-art 的结果，高出 2017 年 Winner
9.4%。
            Abstract. Correlation filter-based trackers heavily suffer from the problem of multiple peaks in their response maps incurred by occlusions.Moreover, the whole tracking pipeline may break down due to the uncertainties brought by shifting among peaks, which will further lead to the degraded correlation filter model. To alleviate the drift problem caused by occlusions, we propose a novel scheme to choose the specific filter model according to different scenarios. Specifically, an eff