GNN-PMB: A Simple but Effective Online 3D Multi-Object Tracker Without Bells and Whistles

Liu, Jianan; Bai, Liping; Xia, Yuxuan; Huang, Tao; Zhu, Bing; Han, Qing‐Long

doi:10.1109/tiv.2022.3217490

Cited by 22 publications

(7 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A second class of methods formulates and solves the MOT problem in the Bayesian estimation framework [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14]. Methods in this class rely on a statistical model for object birth, object motion, and measurement generation [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14]. The statistical model makes it possible to perform a more robust probabilistic "soft" data association and to avoid heuristics for track initialization and termination.…”

Section: A Model-based and Data-driven Motmentioning

confidence: 99%

“…The statistical model makes it possible to perform a more robust probabilistic "soft" data association and to avoid heuristics for track initialization and termination. Methods in this class include the joint probabilistic data association (JPDA) filter [1], the multiple hypothesis tracker (MHT) [36], random finite sets (RFS) filters [9], [10], [11], [12], [13], [14], and belief propagation (BP)-based MOT [3], [4], [5], [6], [7], [8], [9].…”

Section: A Model-based and Data-driven Motmentioning

confidence: 99%

“…It follows the tracking approach of SimpleTrack, except that it never terminates tracks. PMB [13] implements a Poisson multi-Bernoulli filter for MOT that relies on a global nearest neighbor approach for data association. Finally, OGR3MOT [19] utilizes a network flow formulation and transforms the data association problem into a classification problem.…”

Section: Performance Evaluationmentioning

confidence: 99%

“…[6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26] enables emerging applications including autonomous driving, applied ocean sciences, and indoor localization. MOT aims at estimating the states (e.g., positions and possibly other parameters) of moving objects over time, based on measurements provided by sensing technologies such as Light Detection and Ranging (LiDAR), radar, or sonar [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26]. An inherent problem in MOT is measurement-origin uncertainty, i.e., the unknown association between measurements and objects.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Neural Enhanced Belief Propagation for Multiobject Tracking

Liang,

Meyer

2024

IEEE Trans. Signal Process.

View full text Add to dashboard Cite

Algorithmic solutions for multi-object tracking (MOT) are a key enabler for applications in autonomous navigation and applied ocean sciences. State-of-the-art MOT methods fully rely on a statistical model and typically use preprocessed sensor data as measurements. In particular, measurements are produced by a detector that extracts potential object locations from the raw sensor data collected at discrete time steps. This preparatory processing step reduces data flow and computational complexity but may result in a loss of information. Stateof-the-art Bayesian MOT methods that are based on belief propagation (BP) systematically exploit graph structures of the statistical model to reduce computational complexity and improve scalability. However, as a fully model-based approach, BP can provide highly suboptimal estimates when there is a mismatch between the statistical model and the true data-generating process. Existing BP-based MOT methods can further only make use of preprocessed measurements. In this paper, we introduce a variant of BP that combines model-based with data-driven MOT. The proposed neural enhanced belief propagation (NEBP) method complements the statistical model of BP by information learned from raw sensor data. This approach conjectures that the learned information can reduce model mismatch and thus improve data association and false alarm rejection. Our NEBP method improves tracking performance compared to model-based methods. At the same time, it inherits the advantages of BP-based MOT, i.e., it scales only quadratically in the number of objects, and it can thus generate and maintain a large number of object tracks.We evaluate the performance of our NEBP approach for MOT on the nuScenes autonomous driving dataset and demonstrate that it can achieve state-of-the-art performance. In particular, an average multi-object tracking accuracy (AMOTA) of 0.683 was obtained and, compared with non-BP-based methods, identity switches (IDS) and track fragments (Frag) were reduced by 23% and 19%, respectively.

show abstract

Section: A Model-based and Data-driven Motmentioning

confidence: 99%

Section: A Model-based and Data-driven Motmentioning

confidence: 99%

Section: Performance Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Neural Enhanced Belief Propagation for Multiobject Tracking

Liang,

Meyer

2024

IEEE Trans. Signal Process.

View full text Add to dashboard Cite

show abstract

“…In recent years, deep learning methods have been utilized with promising results in image analysis tasks, such as recognition [7], [8], detection [9], [10], [11], segmentation [12], [13], [14], [15], [16], and pose estimation [17]. Many research works on surgical instrument detection have been observed to be applied for identifying and locating surgical instruments intraoperatively, thus achieving surgical instrument tracking [5], [6], [18], surgical skill assessment [19] and surgical instrument monitoring [20].…”

Section: Introductionmentioning

confidence: 99%