Environment-Aware Multi-Target Tracking of Pedestrians

Doellinger, Johannes; Prabhakaran, Vishnu; Fu, Liangcheng; Spies, Markus

doi:10.1109/lra.2019.2898039

Cited by 4 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, it is difficult for twin trackers such as SiamFC to distinguish similar objects, even if the objects have obvious chromatic aberration. This is because SiamFC is insensitive to the underlying characteristics [ 13 ]. By using convolutional network, the tracker can not only obtain semantic information from high-level convolutional features, but also the local features such as texture can be acquired from low-level convolutional features.…”

Section: Target Tracking Methods Based On Deep Learningmentioning

confidence: 99%

Research on Intelligent Target Tracking Algorithm Based on MDNet under Artificial Intelligence

Wang

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Target tracking is an important subject in computer vision technology, which has developed rapidly in recent ten years, and its application have become wider and wider. In this process, it has transferred from a simple experimental tracking environment to a complex real scene where more challenges need to be solved. The rapid development of deep learning has promoted the research progress of digital vision. Target tracking technology is an important foundation of digital vision research, which makes it develop from academia to industry. In this paper, a method of target tracking using MDNet is introduced. Starting with the attention mechanism, two attention mechanisms are added to extract and integrate the better features. Case partitioning is used to reduce the investment of tracking module and minimize the network size during tracking, and its result can be prevented from getting worse. Finally, the experiment is analyzed in detail.

show abstract

Section: Target Tracking Methods Based On Deep Learningmentioning

confidence: 99%

Research on Intelligent Target Tracking Algorithm Based on MDNet under Artificial Intelligence

Wang

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…cle that appears in the surveillance networks. Doellinger et al [9] use tracking methods to predict local statistics about the direction of human motion. A core problem in tracking is how to locate an object accurately and efficiently in challenging scenarios like background clutter, occlusion, scale variation, illumination change, deformation and other variations [42].…”

Section: Introductionmentioning

confidence: 99%

Accurate Anchor Free Tracking

Peng,

Yu,

Wang

et al. 2020

Preprint

View full text Add to dashboard Cite

Visual object tracking is an important application of computer vision. Recently, Siamese based trackers have achieved a good accuracy. However, most of Siamese based trackers are not efficient, as they exhaustively search potential object locations to define anchors and then classify each anchor (i.e., a bounding box). This paper develops the first Anchor Free Siamese Network (AFSN). Specifically, a target object is defined by a bounding box center, tracking offset and object size. All three are regressed by Siamese network with no additional classification or regional proposal, and performed once for each frame. We also tune the stride and receptive field for Siamese network, and further perform ablation experiments to quantitatively illustrate the effectiveness of our AFSN. We evaluate AFSN using five most commonly used benchmarks and compare to the best anchor-based trackers with source codes available for each benchmark. AFSN is 3 × −425× faster than these best anchor based trackers. AFSN is also 5.97% to 12.4% more accurate in terms of all metrics for benchmark sets OTB2015, VOT2015, VOT2016, VOT2018 and Track-ingNet, except that SiamRPN++ is 4% better than AFSN in terms of Expected Average Overlap (EAO) on VOT2018 (but SiamRPN++ is 3.9× slower).

show abstract