BioMed Research International / 2016 / Article / Tab 1 / Research Article
Convolutional Deep Belief Networks for Single-Cell/Object Tracking in Computational Biology and Computer Vision Table 1 A representative success score (AUC) of SRE for different subsets divided based on main variation of the target object. Only the top 5 trackers are displayed for clarity.
Image attributes Ranking The first The second The third The fourth The fifth Fast motion ( ) CDBN-10-2 (0.472) Struck (0.451) TLD (0.385) CXT (0.348) OAB (0.322) Background clutter ( ) CDBN-10-2 (0.414) ASLA (0.410) Struck (0.408) SCM (0.387) VTD (0.377) Motion blur ( ) CDBN-10-2 (0.530) Struck (0.452) TLD (0.392) CXT (0.354) DFT (0.325) Deformation ( ) CDBN-10-2 (0.451) Struck (0.398) ASLA (0.386) DFT (0.364) CPF (0.362) Illumination variation ( ) CDBN-10-2 (0.440) ASLA (0.405) Struck (0.396) SCM (0.389) VTS (0.378) In-plane rotation ( ) CDBN-10-2 (0.422) CXT (0.410) Struck (0.410) ASLA (0.405) SCM (0.399) Low resolution ( ) CDBN-10-2 (0.387) Struck (0.360) MTT (0.326) OAB (0.311) TLD (0.305) Occlusion ( ) CDBN-10-2 (0.441) Struck (0.405) SCM (0.398) TLD (0.384) LSK (0.384) Out-of-plane rotation ( ) CDBN-10-2 (0.427) Struck (0.409) ASLA (0.404) SCM (0.396) VTD (0.392) Out of view ( ) CDBN-10-2 (0.457) Struck (0.421) LOT (0.411) TLD (0.407) CPF (0.394) Scale variation ( ) CDBN-10-2 (0.441) ASLA (0.440) SCM (0.438) Struck (0.395) TLD (0.384)