The Ninth Visual Object Tracking VOT2021 Challenge Results

Kristan, Matej; Matas, Jiřı́; Leonardis, Aleš; Felsberg, Michael; Pflugfelder, Roman; Kämäräinen, Joni-Kristian; Chang, Hyung Jin; Danelljan, Martin; Zajc, Luka Čehovin; Lukežič, Alan; Drbohlav, Ondřej; Käpylä, Jani; Hager, G; Yan, Song; Yang, Jinyu; Zhang, Zhongqun; Fernández, Gustavo J.; Abdelpakey, Mohamed H.; Bhat, Goutam; Çerkezi, Llukman; Çevikalp, Hakan; Chen, Shengyong; Chen, Xin; Cheng, Miao; Cheng, Zimeng; Chiu, Yu‐Chen; Cirakman, Ozgun; Cui, Yutao; Dai, Kenan; Dasari, Mohana Murali; Deng, Qili; Dong, Xingping; Du, Daniel K.; Dunnhofer, Matteo; Feng, Zhen-Hua; Feng, Zhiyong; Fu, Zhihong; Ge, Shiming; Gorthi, Rama Krishna Sai Subrahmanyam; Gu, Yuzhang; Günsel, Bilge; Guo, Qing; Gurkan, Filiz; Han, Wencheng; Huang, Yué; Lawin, Felix Järemo; Jhang, Shang-Jhih; Ji, Rongrong; Jiang, Cheng; Jiang, Yingjie; Juefei-Xu, Felix; Xiao, Ke; Khan, Fahad Shahbaz; Kim, Byeong Hak; Kittler, Josef; Lan, Xiangyuan; Lee, Jun Ha; Leibe, Bastian; Li, Hui; Li, Jianhua; Li, Xianxian; Li, Yuezhou; Liu, Bo; Liu, Chang; Liu, Jingen; Liu, Li; Liu, Qingjie; Lu, Huchuan; Lu, Wei; Luiten, Jonathon; Ma, Jie; Ma, Zhimin; Martinel, Niki; Mayer, Christoph; Memarmoghadam, Alireza; Micheloni, Christian; Niu, Yuzhen; Paudel, Danda Pani; Peng, Houwen; Qiu, Shoumeng; Rajiv, Aravindh; Rana, Muhammad Awais; Robinson, Andreas; Sarıbaş, Hasan; Shao, Ling; Shehata, Mohamed; Shen, Furao; Shen, Jianbing; Simonato, Kristian; Song, Xiaoning; Tang, Zhangyong; Timofte, Radu; Torr, Philip H. S.; Tsai, Chi-Yi; Uzun, Bedirhan; Gool, Luc Van; Voigtlaender, Paul; Wang, Dong; Wang, Guangting; Wang, Liangliang; Wang, Lijun; Wang, Limin; Wang, Linyuan; Wang, Yong; Wang, Yunlong; Wu, Chenyan; Wu, Guorong; Wu, Xiao-Jun; Xie, Fei; Xu, Tianyang; Xu, Xiang; Xue, Wanli; Yan, Bin; Yang, Wankou; Yang, Xiaoyun; Yu, Ye; Yin, Jianqin; Zhang, Chengwei; Zhang, Chunhui; Zhang, Haitao; Zhang, Kaihua; Zhang, Kangkai; Zhang, Xiaohan; Zhang, Xiaolin; Zhang, Xinyu; Zhang, Zhibin; Zhao, Shaochuan; Zhen, Ming Jie; Zhong, Bineng; Zhu, Jiawen; Zhu, Xuefeng

doi:10.1109/iccvw54120.2021.00305

Cited by 72 publications

(26 citation statements)

References 70 publications

(168 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We evaluate our tracker on Visual Object Tracking challenge 2021 (VOT2021) [22] using baseline and real-time experiments. VOT2021 consists of 60 challenging videos with mask annotation.…”

Section: Evaluation On Votmentioning

confidence: 99%

See 1 more Smart Citation

High-Performance Transformer Tracking

Chen¹,

Yan²,

Zhu³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Correlation has a critical role in the tracking field, especially in recent popular Siamese-based trackers. The correlation operation is a simple fusion manner to consider the similarity between the template and the search region. However, the correlation operation is a local linear matching process, losing semantic information and falling into local optimum easily, which may be the bottleneck of designing high-accuracy tracking algorithms. In this work, to determine whether a better feature fusion method exists than correlation, a novel attention-based feature fusion network, inspired by Transformer, is presented. This network effectively combines the template and the search region features using attention. Specifically, the proposed method includes an ego-context augment module based on self-attention and a cross-feature augment module based on cross-attention. First, we present a Transformer tracking (named TransT) method based on the Siamese-like feature extraction backbone, the designed attention-based fusion mechanism, and the classification and regression head. Based on the TransT baseline, we further design a segmentation branch to generate an accurate mask. Finally, we propose a stronger version of TransT by extending TransT with a multi-template design and an IoU prediction head, named TransT-M. Experiments show that our TransT and TransT-M methods achieve promising results on seven popular datasets. Code and models are available at https://github.com/chenxin-dlut/TransT-M.

show abstract

“…We evaluate our tracker on Visual Object Tracking challenge 2021 (VOT2021) [22] using baseline and real-time experiments. VOT2021 consists of 60 challenging videos with mask annotation.…”

Section: Evaluation On Votmentioning

confidence: 99%

“…Studies [19], [20], [21] have predicted the mask and the bounding box. VOT2021 [22] uses mask annotation for evaluation. In this work, we also design a segmentation branch for our tracker to generate an accurate mask while predicting the bounding box by a unified framework.…”

Section: Introductionmentioning

confidence: 99%

High-Performance Transformer Tracking

Chen¹,

Yan²,

Zhu³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Since 2019, there have been 3 VOT-RGBD challenges [17]- [19] held annually, consisting of 13 participating trackers in total. Most participants in VOT just provide a tracking model and brief introduction without specific publications yet they show outstanding performance in multi-challenges.…”

Section: Vot Participantsmentioning

confidence: 99%

“…There are diverse and large datasets for model training and evaluation, such as OTB [11], GOT-10k [12] and TrackingNet [13]. This field is also fueled by the annual VOT challenges [14]- [19].…”

Section: Introductionmentioning

confidence: 99%

“…By providing a test set and evaluation protocols, RGBD tracking can gain more attention in the object tracking community. Until now, there have been 13 participant RGBD trackers in the VOT-RGBD challenges [17]- [19]. According to VOT reports, the VOT trackers show superior performance over traditional solutions, and continuously improve the state-of-the-arts.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

RGBD Object Tracking: An In-depth Review

Yang¹,

Li²,

Yan³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

RGBD object tracking is gaining momentum in computer vision research thanks to the development of depth sensors. Although numerous RGBD trackers have been proposed with promising performance, an in-depth review for comprehensive understanding of this area is lacking. In this paper, we firstly review RGBD object trackers from different perspectives, including RGBD fusion, depth usage, and tracking framework. Then, we summarize the existing datasets and the evaluation metrics. We benchmark a representative set of RGBD trackers, and give detailed analyses based on their performances. Particularly, we are the first to provide depth quality evaluation and analysis of tracking results in depth-friendly scenarios in RGBD tracking. For long-term settings in most RGBD tracking videos, we give an analysis of trackers' performance on handling target disappearance. To enable better understanding of RGBD trackers, we propose robustness evaluation against input perturbations. Finally, we summarize the challenges and provide open directions for this community. All resources are publicly available at https://github.com/memoryunreal/RGBD-tracking-review.

show abstract

Vpit: real-time embedded single object 3D tracking using voxel pseudo images

Oleksiienko,

Nousi,

Passalis

et al. 2024

Neural Comput & Applic

View full text Add to dashboard Cite

In this paper, we propose a novel voxel-based 3D single object tracking (3D SOT) method called Voxel Pseudo Image Tracking (VPIT). VPIT is the first method that uses voxel pseudo images for 3D SOT. The input point cloud is structured by pillar-based voxelization, and the resulting pseudo image is used as an input to a 2D-like Siamese SOT method. The pseudo image is created in the Bird’s-eye View (BEV) coordinates; and therefore, the objects in it have constant size. Thus, only the object rotation can change in the new coordinate system and not the object scale. For this reason, we replace multi-scale search with a multi-rotation search, where differently rotated search regions are compared against a single target representation to predict both position and rotation of the object. Experiments on KITTI [1] Tracking dataset show that VPIT is the fastest 3D SOT method and maintains competitive Success and Precision values. Application of a SOT method in a real-world scenario meets with limitations such as lower computational capabilities of embedded devices and a latency-unforgiving environment, where the method is forced to skip certain data frames if the inference speed is not high enough. We implement a real-time evaluation protocol and show that other methods lose most of their performance on embedded devices; while, VPIT maintains its ability to track the object.

show abstract

The Ninth Visual Object Tracking VOT2021 Challenge Results

Cited by 72 publications

References 70 publications

High-Performance Transformer Tracking

High-Performance Transformer Tracking

RGBD Object Tracking: An In-depth Review

Vpit: real-time embedded single object 3D tracking using voxel pseudo images

Contact Info

Product

Resources

About