Xiantao Hu (胡现韬)
Ph. D. Candidate, Nanjing University of Science and Technology
Address: No. 200, Xiaolingwei Street, Xuanwu District, Nanjing, China
Email: hubulai00 [at] {163.com}

About Me [ Google Scholar ] [ Github ] [ Kaggle ]

I'm a Ph. D. candidate in Nanjing University of Science and Technology(NJUST), advised by Prof. Jian Yang and co-advised by Associate Prof. Ying Tai . I got my master's degree from Guangxi Normal University (GXNU). My master's advisor is Prof. Bineng Zhong.

My recent works are mainly on:
  • Multimodal tracking
  • Visual object tracking

Honor

  • 04/2024 – Outstanding Graduates of General Higher Education in Guangxi Zhuang Autonomous Region
  • 12/2023 – Third in 2023 Yangtze River Delta (Wuhu) Artificial Intelligence Competition - Non motorized vehicle recognition without helmet based on object detection
  • 10/2023 – National Graduate Scholarship
  • 08/2023 – Thrid in IFlytek Developer Competition – Stable Diffusion Identification Challenge
  • 07/2023 – Second in ICCV2023 Workshop challenge – VisDrone2023
  • 06/2023 – Second in CVPR2023 AVA Accessibility Vision and Autonomy Challenge – Segmentation Track
  • 06/2023 – Thrid in CVPR2023 AVA Accessibility Vision and Autonomy Challenge – Keypoint Track
  • 05/2023 – Third in FGVC10 Workshop at CVPR – PlantTraits2023
  • 07/2022 – First in ECV2022 - Outdoor Billboard Recognition

News

  • 04/2024 – One paper accepted by TCSVT
  • 01/2024 – One paper accepted by CSCWD2024
  • 06/2023 – One paper accepted by TCSVT
  • 05/2022 – One paper accepted by Journal of Guangxi Normal University (Natural Science Edition)

Publications

(* indicates equal contribution, # corresponding author)
Transformer Tracking via Frequency Fusion
Xiantao Hu; Bineng Zhong#; Qihua Liang; Shengping Zhang; Ning Li; Xianxian Li; Rongrong Ji
in TCSVT, 2024
[Paper] [BibTex]
In this paper, we propose a frequency fusion based tracker (FFTrack), that achieves better utilization of various frequencies for target tracking by maintaining a relatively balanced frequency.
Towards Modalities Correlation for RGB-T Tracking
Xiantao Hu; Bineng Zhong#; Qihua Liang; Shengping Zhang; Ning Li; Xianxian Li
in TCSVT, 2024
[Paper] [BibTex] [Code]
In this paper, we propose a modalities correlation based RGB-T tracker (MCTrack), that achieves better performance through channel association modeling and fusion of association advantages models.

Review Services

Conference Reviewer
Chinese Conference on Pattern Recognition and Computer Vision (PRCV), 2022,2023

Friendly Links