Modeling local behavior for predicting social interactions towards human tracking

作者：

Highlights：

• We model multiple social effects in pedestrian dynamics.

• We propose a decomposed motion model that approximates complex social interactions.

• The algorithm adjusts the number of basic trackers dynamically based on the exact interaction.

摘要

Highlights•We model multiple social effects in pedestrian dynamics.•We propose a decomposed motion model that approximates complex social interactions.•The algorithm adjusts the number of basic trackers dynamically based on the exact interaction.

论文关键词：Human interaction,Social cues,Interactive Markov Chain Monte Carlo,Multi-object tracking

论文评审过程：Received 7 March 2013, Revised 27 September 2013, Accepted 19 October 2013, Available online 31 October 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.10.019