.png){kind=icon}
返回
SCTracker: Multi-Object Tracking With Shape and Confidence Constraints
.png){kind=icon}
OA
AI
DOI：10.1109/JSEN.2023.3339212.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Detection-based tracking is one of the main methods of multi-object tracking. It can achieve good tracking performance when using excellent detectors but it may associate wrong targets when facing overlapping and low-confidence detections. To address this issue, this article proposes a novel multi-object tracker (SCTracker) by exploiting shape constraint and confidence. In the data association stage, an intersection of union (IoU) distance with shape constraints is developed to calculate the cost matrix between tracks and detections, which can reduce the track of the wrong target with the similar position but inconsistent shape. Moreover, the detection confidence is calculated in the update stage of the Kalman filter to improve the track performance with the inaccurate detection result. Experimental results on the MOT 17 dataset show that the proposed SCTracker can improve the tracking performance of multi-object tracking when compared with the state-of-the-art methods.
Keyword:
Deep learning
distance measurement
motion estimation
object tracking
期刊
IF:
4.5
论文数:
2.1W
被引数:
7.3W
机构
引用论文
.png){kind=icon}
分享收藏
A Simple but Effective Method for Balancing Detection and Re-Identification in Multi-Object Tracking
PREAI
分享收藏
查看更多内容