Skip to main content

Advertisement

Table 2 Qualitative comparison of generative trackers-based human tracking within a camera

From: Human tracking over camera networks: a review

Item no. Used generative trackers Speed Occlusion Scale change Shape deformation
1 Active models-based KF (Jang et al. [4]) High ×
2 Structural KF (Jang et al. [5]) Moderate
3 adaptive KF (Weng et al. [6]) High
4 features-based KF (Li et al. [7]) Moderate
5 MCMC-based PF (Chong et al. [8]) Low
6 Swarm intelligence-based PF (Zhang et al. [9]) Moderate ×
7 Occlusion-aware PF (Meshgi et al. [10]) Low
8 interactive PF (Yang et al. [11]) Low
9 Eigenshape-based KT (Liu et al. [12]) High
10 adaptive MKT (Chu et al. [13]) High
11 fragments-based MKT (Fang et al. [14]) High
12 deformable MKT (Hou et al. [15, 16]) Moderate
  1. Symbols √ and× mean that the used generative trackers-based human tracking within a camera can or cannot deal with the situations of occlusion, scale change, and shape deformation