Discriminative Appearance Modeling with Multi-track Pooling for Real-time Multi-object Tracking

@article{Kim2021DiscriminativeAM,
  title={Discriminative Appearance Modeling with Multi-track Pooling for Real-time Multi-object Tracking},
  author={Chanho Kim and Fuxin Li and Mazen Alotaibi and James M. Rehg},
  journal={2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2021},
  pages={9548-9557}
}
  • Chanho Kim, Fuxin Li, +1 author J. Rehg
  • Published 28 January 2021
  • Computer Science
  • 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
In multi-object tracking, the tracker maintains in its memory the appearance and motion information for each object in the scene. This memory is utilized for finding matches between tracks and detections, and is updated based on the matching. Many approaches model each target in isolation and lack the ability to use all the targets in the scene to jointly update the memory. This can be problematic when there are similarly looking objects in the scene. In this paper, we solve the problem of… 
Online Multi-Object Tracking with Unsupervised Re-Identification Learning and Occlusion Estimation
  • Qiankun Liu, Dongdong Chen, +4 authors Nenghai Yu
  • Computer Science
    Neurocomputing
  • 2022
TLDR
This study shows that, when applied to state-of-the-art MOT methods, the proposed unsupervised re-identification learning is comparable to supervised re-Identification learning, and the tracking performance is further improved by the proposed occlusion estimation module.
MAT: Motion-Aware Multi-Object Tracking
TLDR
This paper proposes an enhanced MOT paradigm, namely Motion-Aware Tracker (MAT), focusing more on various motion patterns of different objects, and introduces the dynamic reconnection context module, which aims to balance the robustness of long-range motion-based reconnection.
DeepScale: An Online Frame Size Adaptation Framework to Accelerate Visual Multi-object Tracking
TLDR
DeepScale is proposed, a model agnostic frame size selection approach that operates on top of existing fully convolutional network-based trackers to accelerate tracking throughput and find a suitable trade-off between tracking accuracy and speed by adapting frame sizes at run time.
DeepScale: An Online Frame Size Adaptation Approach to Accelerate Visual Multi-object Tracking
TLDR
DeepScale is proposed, a model agnostic frame size selection approach that operates on top of existing fully convolutional network-based trackers to accelerate tracking throughput and incorporate detectability scores into a one-shot tracker architecture.
Real-time Generalized Sensor Fusion with Transformers
3D Multi Object Tracking (MOT) is essential for the safe deployment of self-driving vehicles. While major progress has been made in 3D object detection and machinelearned tracking approaches, real
TransCenter: Transformers with Dense Queries for Multiple-Object Tracking
TLDR
This work proposes TransCenter, the first transformer-based architecture for tracking the centers of multiple targets, using the use of dense queries in a double-decoder network, to be able to robustly infer the heatmap of targets’ centers and associate them through time.
STURE: Spatial-Temporal Mutual Representation Learning for Robust Data Association in Online Multi-Object Tracking
  • Haidong Wang, Zhiyong Li, Yaping Li, Ke Nai, Ming Wen
  • Computer Science
  • 2022
Online multi-object tracking (MOT) is a longstanding task for computer vision and intelligent vehicle platform. At present, the main paradigm is tracking-by-detection, and the main difficulty of this
DeepScale: Online Frame Size Adaptation for Multi-object Tracking on Smart Cameras and Edge Servers
TLDR
This work proposes DeepScale, a model agnostic frame size selection approach that operates on top of existing fully convolutional network-based trackers to accelerate tracking throughput and proposes two computation partition schemes tailored for MOT.

References

SHOWING 1-10 OF 50 REFERENCES
Online Multi-object Tracking Using CNN-Based Single Object Tracker with Spatial-Temporal Attention Mechanism
TLDR
A CNN-based framework for online MOT that utilizes the merits of single object trackers in adapting appearance models and searching for target in the next frame and introduces spatial-temporal attention mechanism (STAM) to handle the drift caused by occlusion and interaction among targets.
Real-Time Multiple People Tracking with Deeply Learned Candidate Selection and Person Re-Identification
TLDR
This paper proposes to handle unreliable detection by collecting candidates from outputs of both detection and tracking, and adopts a deeply learned appearance representation, which is trained on large-scale person re-identification datasets, to improve the identification ability of the tracker.
Online Multi-Object Tracking with Dual Matching Attention Networks
TLDR
This paper introduces a cost-sensitive tracking loss based on the state-of-the-art visual tracker which encourages the model to focus on hard negative distractors during online learning and proposes Dual Matching Attention Networks (DMAN) with both spatial and temporal attention mechanisms.
Learning Discriminative Appearance Models for Online Multi-Object Tracking With Appearance Discriminability Measures
TLDR
This paper proposes a discriminative online appearance learning using a partial least square (PLS) method and evaluates the performance of the MOT method on public MOT benchmark challenge datasets and shows the competitive performance compared to other state-of-the-art batch and online tracking methods.
Multi-object Tracking with Neural Gating Using Bilinear LSTM
TLDR
A novel recurrent network model, the Bilinear LSTM, is proposed in order to improve the learning of long-term appearance models via a recurrent network based on intuitions drawn from recursive least squares.
Online Multi-Object Tracking With Instance-Aware Tracker and Dynamic Model Refreshment
TLDR
This paper proposes an instance-aware tracker to integrate SOT techniques for MOT by encoding awareness both within and between target models, and considers response maps from all target models and assigns spatial locations exclusively to optimize the overall accuracy.
Online Multi-Object Tracking with Historical Appearance Matching and Scene Adaptive Detection Filtering
TLDR
The historical appearance matching method and joint-input siamese network which was trained by 2-step process can prevent tracking failures although objects are temporally occluded or last matching information is unreliable.
Customized Multi-person Tracker
TLDR
A hierarchical correlation clustering (HCC) framework is introduced, in which it is shown that a few high-quality training examples can be used to fine-tune pre-trained CNNs, thereby teaching them to recognize the uniqueness of people’s appearance in the test sequence.
Recurrent Autoregressive Networks for Online Multi-object Tracking
TLDR
This work proposes the Recurrent Autoregressive Network (RAN), a temporal generative modeling framework to characterize the appearance and motion dynamics of multiple objects over time and achieves top-ranked results on the two benchmarks.
How to Train Your Deep Multi-Object Tracker
TLDR
A differentiable proxy of MOTA and MOTP is proposed, which is combined in a loss function suitable for end-to-end training of deep multi-object trackers and establishes a new state of the art on the MOTChallenge benchmark.
...
1
2
3
4
5
...