How to Train Your Deep Multi-Object Tracker

Xu, Yong; sep, Aljosa; Ban, Yutong; Horaud, Radu; Leal-Taixé, Laura; Alameda-Pineda, Xavier

doi:10.1109/cvpr42600.2020.00682

Cited by 187 publications

(106 citation statements)

References 36 publications

Supporting

Mentioning

106

Contrasting

Order By: Relevance

“…Among all the strategies of SLAM, the visual SLAM (vSLAM), which aims to structure the motion of the agents using only images captured by the camera, is one of the most popular approaches since it is inexpensive and common in realistic applications. Recent trends in multiobject tracking [5][6][7] and semantic mapping [8,9] extend the capacity of sensing and understanding for SLAM systems, which significantly benefits some key emerging applications like safety monitoring and autonomous driving. Although the tracking and mapping tasks in SLAM can calibrate each other in a comple-mentary mechanism, almost every vSLAM system suffers from accumulating errors coming from motion blur and illumination variance, and the most effective method to re-locate itself using images only is the loop closure detection [10], which is the main topic of this paper.…”

Section: Introductionmentioning

confidence: 99%

PLSAV: Parallel loop searching and verifying for loop closure detection

Yang

Pan

Deng

et al. 2021

IET Intelligent Trans Sys

View full text Add to dashboard Cite

Visual simultaneous localization and mapping (vSLAM), one of the most important applications in autonomous vehicles and robots to estimate the position and pose using inexpensive visual sensors, suffers from error accumulation for long-term navigation without loop closure detection. Recently, deep neural networks (DNNs) are leveraged to achieve high accuracy for loop closure detection, however the execution time is much slower than those employing handcrafted visual features. In this paper, a parallel loop searching and verifying method for loop closure detection with both high accuracy and high speed, which combines two parallel tasks using handcrafted and DNN features, respectively, is proposed. A fast loop searching is proposed to link the bag-of-words features and histogram for higher accuracy, and it splits the images into multiple grids for high parallelism; meanwhile, a DNN feature extractor is utilized for further verification. A loop state control method based on a finite state machine to control these tasks is designed, wherein the loop closure detection is described as a context-related procedure. The framework is implemented on a real machine, and the top-2 best accuracy and fastest execution time of 80-543 frames per second (min: 1.84ms, and max: 12.45ms) are achieved on several public benchmarks compared with some existing algorithms. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Section: Introductionmentioning

confidence: 99%

PLSAV: Parallel loop searching and verifying for loop closure detection

Yang

Pan

Deng

et al. 2021

IET Intelligent Trans Sys

View full text Add to dashboard Cite

show abstract

“…The authors of some of the methods available in the latest MOT have not released the source code [39][40][41][42][43][44][45][46][47][48]. Some of the available methods were compiled but unfortunately required additional code modifications, input format adjustments, or generated erroneous results [49][50][51][52].…”

Section: Discussionmentioning

confidence: 99%

Application of Crowd Simulations in the Evaluation of Tracking Algorithms

Staniszewski

Foszner

Kostorz

et al. 2020

Sensors

View full text Add to dashboard Cite

Tracking and action-recognition algorithms are currently widely used in video surveillance, monitoring urban activities and in many other areas. Their development highly relies on benchmarking scenarios, which enable reliable evaluations/improvements of their efficiencies. Presently, benchmarking methods for tracking and action-recognition algorithms rely on manual annotation of video databases, prone to human errors, limited in size and time-consuming. Here, using gained experiences, an alternative benchmarking solution is presented, which employs methods and tools obtained from the computer-game domain to create simulated video data with automatic annotations. Presented approach highly outperforms existing solutions in the size of the data and variety of annotations possible to create. With proposed system, a potential user can generate a sequence of random images involving different times of day, weather conditions, and scenes for use in tracking evaluation. In the design of the proposed tool, the concept of crowd simulation is used and developed. The system is validated by comparisons to existing methods.

show abstract

“…e study of Xu et al [55] is an improved version of [54], and the former has more experiments than the latter. An end-to-end training framework for MOT methods is proposed, which can train the evaluation indicators of MOT.…”

Section: Mathematical Problems In Engineeringmentioning

confidence: 99%

“…An end-to-end training framework for MOT methods is proposed, which can train the evaluation indicators of MOT. Most of the methods in [55] are derived from [54]. e final experimental results show that after using the deepMOT training framework to train the tracker, the performance of the tracking method can be improved to a certain extent.…”

Section: Mathematical Problems In Engineeringmentioning

confidence: 99%

Data Association Methods via Video Signal Processing in Imperfect Tracking Scenarios: A Review and Evaluation

Liu

Lin

et al. 2020

Mathematical Problems in Engineering

View full text Add to dashboard Cite

In 5G scenarios, there are a large number of video signals that need to be processed. Multiobject tracking is one of the main directions in video signal processing. Data association is a very important link in tracking algorithms. Complexity and efficiency of association method have a direct impact on the performance of multiobject tracking. Breakthroughs have been made in data association methods based on deep learning, and the performance has been greatly improved compared with traditional methods. However, there is a lack of overviews about data association methods. Therefore, this article first analyzes characteristics and performance of three traditional data association methods and then focuses on data association methods based on deep learning, which is divided into different deep network structures: SOT methods, end-to-end methods, and Wasserstein metric methods. The performance of each tracking method is compared and analyzed. Finally, it summarizes the current common datasets and evaluation criteria for multiobject tracking and discusses challenges and development trends of data association technology and data association methods which ensure robust and real time need to be continuously improved.

show abstract

How to Train Your Deep Multi-Object Tracker

Cited by 187 publications

References 36 publications

PLSAV: Parallel loop searching and verifying for loop closure detection

PLSAV: Parallel loop searching and verifying for loop closure detection

Application of Crowd Simulations in the Evaluation of Tracking Algorithms

Data Association Methods via Video Signal Processing in Imperfect Tracking Scenarios: A Review and Evaluation

Contact Info

Product

Resources

About