Markov chain Monte Carlo based video tracking algorithm
- Authors: Kuplyakov D.1, Shalnov E.1, Konushin A.1,2
-
Affiliations:
- Moscow State University Department of Computational Mathematics and Cybernetics
- NRU Higher School of Economics
- Issue: Vol 43, No 4 (2017)
- Pages: 224-229
- Section: Article
- URL: https://journals.rcsi.science/0361-7688/article/view/176517
- DOI: https://doi.org/10.1134/S0361768817040053
- ID: 176517
Cite item
Abstract
The paper considers a problem of multiple person tracking. We present the algorithm to automatic people tracking on surveillance videos recorded by static cameras. Proposed algorithm is an extension of approach based on tracking-by-detection of people heads and data association using Markov chain Monte Carlo (MCMC). Short track fragments (tracklets) are built by local tracking of people heads. Tracklet postprocessing and accurate results interpolation were shown to reduce number of false positives. We use position deviations of tracklets and revised entry/exit points factor to separate pedestrians from false positives. The paper presents a new method to estimate body position, that increases precision of tracker. Finally, we switched HOG-based detector to cascade one. Our evaluation shows proposed modifications significantly increase tracking quality.
About the authors
D. Kuplyakov
Moscow State University Department of Computational Mathematics and Cybernetics
Author for correspondence.
Email: denis.kuplyakov@graphics.cs.msu.ru
Russian Federation, Moscow, 119991
E. Shalnov
Moscow State University Department of Computational Mathematics and Cybernetics
Email: denis.kuplyakov@graphics.cs.msu.ru
Russian Federation, Moscow, 119991
A. Konushin
Moscow State University Department of Computational Mathematics and Cybernetics; NRU Higher School of Economics
Email: denis.kuplyakov@graphics.cs.msu.ru
Russian Federation, Moscow, 119991; Moscow, 101000