AutoTrajectory: Label-Free Trajectory Extraction and Prediction from Videos Using Dynamic Points

Ma, Yuexin; Zhu, Xinge; Cheng, Xinjing; Yang, Ruigang; Liu, Jiming; Manocha, Dinesh

doi:10.1007/978-3-030-58601-0_38

Cited by 13 publications

(3 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In both works, detection and tracking errors were treated as a minor nuisance to demonstrate the robustness of the proposed method. Recently, however, it is argued in [19] that extracting human trajectories from raw videos remains a challenging problem. It further leveraged unlabeled videos for training deep prediction models.…”

Section: A Human Trajectory Predictionmentioning

confidence: 99%

See 1 more Smart Citation

Towards Robust Human Trajectory Prediction in Raw Videos

Yu¹,

Zhou

2021

Preprint

View full text Add to dashboard Cite

Human trajectory prediction has received increased attention lately due to its importance in applications such as autonomous vehicles and indoor robots. However, most existing methods make predictions based on human-labeled trajectories and ignore the errors and noises in detection and tracking. In this paper, we study the problem of human trajectory forecasting in raw videos, and show that the prediction accuracy can be severely affected by various types of tracking errors. Accordingly, we propose a simple yet effective strategy to correct the tracking failures by enforcing prediction consistency over time. The proposed "re-tracking" algorithm can be applied to any existing tracking and prediction pipelines. Experiments on public benchmark datasets demonstrate that the proposed method can improve both tracking and prediction performance in challenging real-world scenarios. The code and data are available at https://git.io/retracking-prediction.

show abstract

Section: A Human Trajectory Predictionmentioning

confidence: 99%

“…Baseline and implement details. It is non-trivial to detect small objects from bird's-eye view [19] that the state-ofthe-art object detectors [39] failed on SDD. In this study, a motion detector [40] is adopted for SDD but does not perform well in challenging situations such as shaky camera and crowded area.…”

Section: A Algorithmmentioning

confidence: 99%

Towards Robust Human Trajectory Prediction in Raw Videos

Yu¹,

Zhou

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The topic of human behavior prediction is of great interest to robotics and computer vision communities. Behavior prediction has been used in many applications such as humanobject [11], [12] and human-human interaction [13], [14], risk assessment [15], [16], anomaly detection [17], surveillance [18], [7], sports forecasting [19], [20] and intelligent driving systems [21], [5].…”

Section: A Behavior Predictionmentioning

confidence: 99%

Multi-Modal Hybrid Architecture for Pedestrian Action Prediction

Rasouli¹,

Yau²,

Rohani³

et al. 2020

Preprint

View full text Add to dashboard Cite

Pedestrian behavior prediction is one of the major challenges for intelligent driving systems in urban environments. Pedestrians often exhibit a wide range of behaviors and adequate interpretations of those depend on various sources of information such as pedestrian appearance, states of other road users, the environment layout, etc. To address this problem, we propose a novel multi-modal prediction algorithm that incorporates different sources of information captured from the environment to predict future crossing actions of pedestrians. The proposed model benefits from a hybrid learning architecture consisting of feedforward and recurrent networks for analyzing visual features of the environment and dynamics of the scene. Using the existing 2D pedestrian behavior benchmarks and a newly annotated 3D driving dataset, we show that our proposed model achieves state-of-the-art performance in pedestrian crossing prediction.

show abstract

A Method for Partitioning the Topological Structure of Groups Based on Hypergraphs

Wang,

2023

Lecture Notes in Electrical Engineering

View full text Add to dashboard Cite

AutoTrajectory: Label-Free Trajectory Extraction and Prediction from Videos Using Dynamic Points

Cited by 13 publications

References 33 publications

Towards Robust Human Trajectory Prediction in Raw Videos

Towards Robust Human Trajectory Prediction in Raw Videos

Multi-Modal Hybrid Architecture for Pedestrian Action Prediction

A Method for Partitioning the Topological Structure of Groups Based on Hypergraphs

Contact Info

Product

Resources

About