ByteTrack: Multi-object Tracking by Associating Every Detection Box

Zhang, Yifu; Sun, Peize; Jiang, Yi; Yu, Dongdong; Yuan, Zehuan; Luo, Ping; Liu, Wenyu; Wang, Xinggang

doi:10.1007/978-3-031-20047-2_1

Cited by 732 publications

(427 citation statements)

References 62 publications

Supporting

Mentioning

274

Contrasting

Order By: Relevance

“…We also evaluated a few recent state-of-the-art models on the PersonPath22 dataset, including 1), zero-shot IdFree [52] model in which the embedding component is trained without any person identity annotation, 2), TrackFormer [41] whose underlying detector is transformer based Detr [10], as well as 3), Byte-Track [67] that is based on the state-of-the-art singe-tage YOLOX detector [28]. As can be clearly seen, ByteTrack achieves the best MOTA and IDF1, and the zero-shot IDFree model outperforms most recent state-of-the-art tracking models even though it is not trained on the target PersonPath22 dataset.…”

Section: Model Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

Large Scale Real-World Multi-person Tracking

Shuai

Bergamo

Büchler

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

This paper presents a new large scale multi-person tracking dataset -PersonPath22, which is over an order of magnitude larger than currently available high quality multi-object tracking datasets such as MOT17, HiEve, and MOT20 datasets. The lack of large scale training and test data for this task has limited the community's ability to understand the performance of their tracking systems on a wide range of scenarios and conditions such as variations in person density, actions being performed, weather, and time of day. PersonPath22 dataset was specifically sourced to provide a wide variety of these conditions and our annotations include rich meta-data such that the performance of a tracker can be evaluated along these different dimensions. The lack of training data has also limited the ability to perform end-to-end training of tracking systems. As such, the highest performing tracking systems all rely on strong detectors trained on external image datasets. We hope that the release of this dataset will enable new lines of research that take advantage of large scale video based training data.

show abstract

Section: Model Evaluationmentioning

confidence: 99%

“…ByteTrack [67]. For ByteTrack, the detector is YOLOX [28] with YOLOX-X as the backbone and COCO-pretrained model as the initialized weights.…”

Section: D2 Implementation Detailsmentioning

confidence: 99%

Large Scale Real-World Multi-person Tracking

Shuai

Bergamo

Büchler

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…A sample image and its corresponding annotation are shown in Figure 2A. The dataset consists of eight parts, with Four state-of-the-art MOT methods are tested on the proposed dataset, which are ByteTrack (Zhang et al, 2021a), ByteTrack with NSA Kalman filter (Du et al, 2021), FairMOT (Zhang et al, 2021b), and SORT (Bewley et al, 2016) 2 . They are finetuned on our dataset using their default hyperparameters.…”

Section: Image Annotation and Dataset Constructionmentioning

confidence: 99%

LettuceMOT: A dataset of lettuce detection and tracking with re-identification of re-occurred plants for agricultural robots

Wang

et al. 2022

Front. Plant Sci.

View full text Add to dashboard Cite

show abstract

“…All raw videos were recorded in outdoor scenarios. For each video, we first performed ByteTrack [48] to generate human bounding boxes with unified IDs, and used them to crop out RGB sequences of each identity, which is then split into several short sequences (about 200 frames). After that, we merged the sequences of the same identity from different videos and manually labeled the clothes IDs.…”

Section: The Rccvreid Datasetmentioning

confidence: 99%

A benchmark for clothes variation in person re‐identification

Wang

Chen

et al. 2020

Int J Intell Syst

View full text Add to dashboard Cite

Person re-identification (re-ID) has drawn attention significantly in the computer vision society due to its application and research significance. It aims to retrieve a person of interest across different camera views. However, there are still several factors that hinder the applications of person re-ID. In fact, most common data sets either assume that pedestrians do not change their clothing across different camera views or are taken under constrained environments. Those constraints simplify the person re-ID task and contribute to early development of person re-ID, yet a person has a great possibility to change clothes in real life. To facilitate the research toward conquering those issues, this paper mainly introduces a new benchmark data set for person reidentification. To the best of our knowledge, this data set is currently the most diverse for person re-identification. It contains 107 persons with 9,738 images, captured in 15 indoor/outdoor scenes from September 2019 to December 2019, varying according to viewpoints, lighting, resolutions, human pose, seasons, backgrounds, and clothes especially. We hope that this benchmark data set will encourage further research on person re-identification with clothes variation. Moreover, we also perform extensive analyses on this data set using several

show abstract

ByteTrack: Multi-object Tracking by Associating Every Detection Box

Cited by 732 publications

References 62 publications

Large Scale Real-World Multi-person Tracking

Large Scale Real-World Multi-person Tracking

LettuceMOT: A dataset of lettuce detection and tracking with re-identification of re-occurred plants for agricultural robots

A benchmark for clothes variation in person re‐identification

Contact Info

Product

Resources

About