Probabilistic 3D Multi-Modal, Multi-Object Tracking for Autonomous Driving

Chiu, Hsu-kuang; Li, Jie; Ambruş, Rareş; Bohg, Jeannette

doi:10.1109/icra48506.2021.9561754

Cited by 125 publications

(64 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, the tracking method proposed in [7] uses a heuristic to create tracks and a greedy matching algorithm based on the Euclidean distance to associate CenterPoint object detections to tracks. Chiu et al [14] follows a similar strategy but makes use of a hybrid distance that combines the Mahalanobis distance and the so-called deep feature distance. Finally, the method introduced by Zaech et al [15] utilizes a network flow formulation and transforms the DA problem into a classification problem.…”

Section: B Performance Evaluationmentioning

confidence: 99%

Neural Enhanced Belief Propagation for Data Association in Multiobject Tracking

Liang¹,

Meyer²

2022

Preprint

View full text Add to dashboard Cite

Situation-aware technologies enabled by multiobject tracking (MOT) methods will create new services and applications in fields such as autonomous navigation and applied ocean sciences. Belief propagation (BP) is a state-of-the-art method for Bayesian MOT but fully relies on a statistical model and preprocessed sensor measurements. In this paper, we establish a hybrid method for model-based and data-driven MOT. The proposed neural enhanced belief propagation (NEBP) approach complements BP by information learned from raw sensor data with the goal to improve data association and to reject false alarm measurements. We evaluate the performance of our NEBP approach for MOT on the nuScenes autonomous driving dataset and demonstrate that it can outperform state-ofthe-art reference methods.

show abstract

Section: B Performance Evaluationmentioning

confidence: 99%

Neural Enhanced Belief Propagation for Data Association in Multiobject Tracking

Liang¹,

Meyer²

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…For example, CenterPoint [1] achieved this by estimating the velocity of each object. Other methods [16], [2] exploited appearance information from camera to enhance the association quality. SimTrack [17] avoided heuristic matching step by joint detection and tracking in an end-to-end manner.…”

Section: B 3d Multi-object Trackingmentioning

confidence: 99%

Learning Moving-Object Tracking with FMCW LiDAR

Gu¹,

Cheng²,

Wang³

et al. 2022

Preprint

View full text Add to dashboard Cite

In this paper, we propose a learning-based moving-object tracking method utilizing our newly developed LiDAR sensor, Frequency Modulated Continuous Wave (FMCW) LiDAR. Compared with most existing commercial LiDAR sensors, our FMCW LiDAR can provide additional Doppler velocity information to each 3D point of the point clouds. Benefiting from this, we can generate instance labels as ground truth in a semi-automatic manner. Given the labels, we propose a contrastive learning framework, which pulls together the features from the same instance in embedding space and pushes apart the features from different instances, to improve the tracking quality. Extensive experiments are conducted on our recorded driving data, and the results show that our method outperforms the baseline methods by a large margin.

show abstract

“…In addition to bounding box information, multiple works include 2D/3D appearance features to substantiate pair-wise affinity representation [23]. Deep learning allows to learn semantic features via encoding: Popular methods [12], [22] use image classification networks as encoders for representing image data and PointNet [24] architectures for learning point cloud features.…”

Section: Related Workmentioning

confidence: 99%

“…While the detections are often derived only from a single sensor modality such as LiDAR or camera, the entirety of modalities can still be utilized for improved similarity finding of detections in the tracking task. As discussed in multiple prior works [22], [23], [26] sensor fusion is still an ongoing research matter. Our approach fuses 3D pose & motion features (3D-PM) from bounding boxes with 2D as well as 3D appearance features from (surround) cameras (2D-A), LiDAR (3D-AL) as well as radar sensors (3D-AR).…”

Section: A Feature Representationmentioning

confidence: 99%

See 1 more Smart Citation

3D Multi-Object Tracking Using Graph Neural Networks with Cross-Edge Modality Attention

Buchner¹,

Valada²

2022

Preprint

View full text Add to dashboard Cite

Online 3D multi-object tracking (MOT) has witnessed significant research interest in recent years, largely driven by demand from the autonomous systems community. However, 3D offline MOT is relatively less explored. Labeling 3D trajectory scene data at a large scale while not relying on high-cost human experts is still an open research question. In this work, we propose Batch3DMOT that follows the tracking-by-detection paradigm and represents real-world scenes as directed, acyclic, and category-disjoint tracking graphs that are attributed using various modalities such as camera, LiDAR, and radar. We present a multi-modal graph neural network that uses a cross-edge attention mechanism mitigating modality intermittence, which translates into sparsity in the graph domain. Additionally, we present attention-weighted convolutions over frame-wise k-NN neighborhoods as suitable means to allow information exchange across disconnected graph components. We evaluate our approach using various sensor modalities and model configurations on the challenging nuScenes and KITTI datasets. Extensive experiments demonstrate that our proposed approach yields an overall improvement of 2.8% in the AMOTA score on nuScenes thereby setting a new benchmark for 3D tracking methods and successfully enhances false positive filtering.

show abstract

Probabilistic 3D Multi-Modal, Multi-Object Tracking for Autonomous Driving

Cited by 125 publications

References 27 publications

Neural Enhanced Belief Propagation for Data Association in Multiobject Tracking

Neural Enhanced Belief Propagation for Data Association in Multiobject Tracking

Learning Moving-Object Tracking with FMCW LiDAR

3D Multi-Object Tracking Using Graph Neural Networks with Cross-Edge Modality Attention

Contact Info

Product

Resources

About