Xin Wang

Xin Wang

A CV&ML Learner

About me

I am Xin Wang (王鑫 in Chinese), I received my doctoral degree from Beijing Jiaotong University, advised by Prof. Tangwen Yang and Prof. Qiuqi Ruan. Before that, I received my master’s degree from Shenzhen University, advised by Prof. Weixin Xie. My research interest lies at the intersection of Computer Vision and Deep Learning. In particular, my work focuses on multimodal large language model, embodied intelligence, crowd behavior analysis, self/semi-supervised learning, and continual domain adaptation.

Research Interests

Currently, my research mainly centers on

  • Multimodal large language model: Vision-Language large model, Efficient fine-tuning, Model data leakage.
  • Crowd behavior analysis: Crowd counting and localization, abnormal behavior detection.
  • Data-efficient learning: Semi-supervised learning (SSL), Few-shot learning.
  • Continual learning: Continual unsupervised domain adaptation(CUDA), Test-time adaptation(TTA).

Selected Publications

X. Wang, T. Yang and Q. Ruan et.al. " Hybrid Perturbation Strategy for Semi-supervised Crowd Counting ". IEEE Transactions on Image Processing(TIP), 2024. (CCF-A, SCI, IF=10.6) [paper]

Y. Zhan, X. Wang, T. Yang and Q. Ruan et.al. " TG-Pose: Delving into Topology and Geometry for Category-level Object Pose Estimation. ". IEEE Transactions on Multimedia (TMM), 2024. (CCF-B, SCI, IF=7.8) [project].

X. Wang, Y. Zhan, Y. Zhao, T. Yang and Q. Ruan. " Semi-Supervised Crowd Counting With Spatial Temporal Consistency and Pseudo-Label Filter ". IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), 2023.(CCF-B, SCI, IF=8.4) [paper]

X. Wang, T. Yang and Q. Ruan et.al. " DPCA: Density Prototype Contrast Adaptation for Continual Unsupervised Domain Adaptation Crowd Counting. ". IEEE Transactions on Multimedia, 2024.

X. Wang, T. Yang and Q. Ruan et.al " Multi-scale context aggregation network with attention-guided for crowd counting ". 15th IEEE International Conference on Signal Processing, 2020.(Best paper) [paper] [code]

X. Wang, W. Xie et.al. " Learning spatiotemporal features with 3DCNN and ConvGRU for video anomaly detection ". 14th IEEE International Conference on Signal Processing, 2018. [paper]