A Continuous, Full-scope, Spatio-temporal Tracking Metric based on KL-divergence

被引:2
作者
Adams, Terry [1 ]
机构
[1] US Govt, Washington, DC 20535 USA
来源
2019 IEEE WINTER APPLICATIONS OF COMPUTER VISION WORKSHOPS (WACVW) | 2019年
关键词
PERFORMANCE EVALUATION; INFORMATION-THEORY;
D O I
10.1109/WACVW.2019.00010
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
A unified metric is given for the evaluation of object tracking systems. The metric is inspired by KL-divergence or relative entropy, which is commonly used to evaluate clustering techniques. Since tracking problems are,fundamentally different from clustering, the components of KL-divergence are recast to handle various types of tracking errors (i.e., false alarms, missed detections, merges, splits). Scoring results are given on a standard tracking dataset (Oxford Town Centre Dataset), as well as several simulated scenarios. Also, this new metric is compared with several other metrics including the commonly used Multiple Object Tracking Accuracy metric. In the final section, advantages of this metric are given including the fact that it is continuous, parameter-less and comprehensive.
引用
收藏
页码:16 / 24
页数:9
相关论文
共 26 条
[1]  
[Anonymous], 2015, IEEE C COMP VIS PATT
[2]  
[Anonymous], IEEE C COMP VIS PATT
[3]  
[Anonymous], UA DETRAC NEW BENCHM
[4]  
[Anonymous], 2009 WORKSH APPL COM
[5]  
[Anonymous], IEEE INT WORKSH PERF
[6]  
[Anonymous], AIPR
[7]  
[Anonymous], IEEE INT WORKSH PERF
[8]  
[Anonymous], IEEE 12 INT C COMP V
[9]  
[Anonymous], IEEE INT WORKSH PERF
[10]  
[Anonymous], IEEE INT WORKSH PERF