Research‎ > ‎Image Analysis‎ > ‎

Long-term Object tracking

물체 추적이란 첫 프레임에서 물체의 위치가 주어지면, 그 이후 나머지 프레임들에서 물체의 위치를 추정하는 것을 의미합니다. 이 분야는 단기적 물체 추적과 장기적 물체 추적으로 나눌 수 있습니다. 영상 내에서 물체가 존재한다는 제한적인 가정을 가지고 있는 단기적 물체 추적과 다르게, 장기적 물체 추적은 물체가 영상 내에서 사라졌다가 다시 나타난다고 하더라도 물체의 위치를 다시 추정하는 재검출 과정이 포함되어 있습니다. 따라서 장기적 물체 추적은 보다 일반적인 물체 추적 방식이며 오랜 시간동안 물체 추적이 가능하기 때문에 군사적 응용에 적합합니다.



A Memory Model based on the Siamese Network for Long-term Tracking
    (MMLT) *3rd prize winner at the VOT-LT2018 challenge* [Github]


Hankyeol Lee, Seokeon Choi, and Changick Kim

  • Abstract

We propose a novel memory model using deep convolutional features for long-term tracking to handle the challenging issues, including visual deformation or target disappearance. Our memory model is separated into short- and long-term stores inspired by Atkinson-Shiffrin Memory Model (ASMM). In the tracking step, the bounding box of the target is estimated by the Siamese features obtained from both memory stores to accommodate changes in the visual appearance of the target. In the re-detection step, we take features only in the long-term store to alleviate the drift problem. At this time, we adopt a coarse-to-fine strategy to detect the target in the entire image without the dependency of the previous position. In the end, we employ Regional Maximum Activation of Convolutions (R-MAC) as key criteria. Our tracker achieves an F-score of 0.52 on the LTB35 dataset, which is 0.04 higher than the performance of the state-of-the-art algorithm.


  • Demonstration video (MMLT) 

‎(ECCV 2018 Workshop)‎ Long-term tracking demo ‎(MMLT)‎




  • "Short-term tracking" vs "Long-term tracking"





  • Overall flowchart




  • VOT-LT2018 benchmark




  • Ablation study





  • Qualitative results




  • Hankyeol Lee*, Seokeon Choi*, and Changick Kim, "A Memory Model based on the Siamese Network for Long-term Tracking," in Proc. European Conference on Computer Vision Workshop (ECCVW),  Munich, Germany, Sep. 8-14, 2018. (* These two authors contributed equally)