Chengbo Dong

About Me

I received a MS degree in Computer Science at Renmin University of China, advied by Prof. Xirong Li and Prof. Jieping Xu.
I received a BE degree in Automation at Beihang University.
My research interests are Computer Vision, Media Forensics and Cross-Modal Computing.
Contact me at cb_dong@foxmail.com

Chengbo Dong*; Xinru Chen*; Ruohan Hu; Juan Cao; Xirong Li. “MVSS-Net: Image Manipulation Detection by Multi-View Multi-Scale Supervision.” In IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022.paper.
Xinru Chen*; Chengbo Dong*; Jiaqi Ji; Juan Cao; Xirong Li. “Image Manipulation Detection by Multi-View Multi-Scale Supervision.” In ICCV , 2021.(co-author). [paper] [poster] [code] [YouTube] [中文速览]
Chengbo Dong*; Xinru Chen*; Aozhu Chen; Fan Hu; Zihan Wang; Xirong Li. “Multi-Level Visual Representation with Semantic-Reinforced Learning for Video Captioning.” In ACM Multimedia (Grand Challenge Track), 2021. (co-author) [paper] [YouTube]
Aozhu Chen; Ziyuan Wang; Chengbo Dong; Kaibin Tian; Ruixiang Zhao; Xun Liang; Zhanhui Kang; Xirong Li. “ChinaOpen: A Dataset for Open-world Multimodal Learning.” In ACM Multimedia, 2023. [paper] [page][code]

Graduate Student, 2020.09 - 2023.06

Cross-modal computing
- Video Captioning & Video-Text Retrieval
  - Two papers accepted by ACMMM’21 & 23.
  - ACMMM’21 Grand Challenge: Pretraining for video captioning, top 3.
  - 2021 *TREC Video Retrieval Evaluation (TRECVID) *VTT top 2, AVS top3.
Media forencis
- Image Manipulation Forensics：Classification and segmentation on manipulated images.
  - One conference paper accepted by ICCV’21.
  - One journal paper accepted by T-PAMI’22
  - Security AI Challenge: Forgery Detection on Certificate Image, Alibaba, top11
- DeepFake Detection：Classification on authentic/deepfake-tampered/GAN-generated human faces
  - Faceforensics Benchmark, top20
  - 2020 China Artificial Competition, Group B, top2
  - A visiting student in Cross-Modal Computing Lab at ICT, CAS

Software Development Engineer Intern, 2022.05 - 2022.10

Research Intern, 2021.12 - 2021.03

Restoration on Manipulated Images against Social Media Transmission
- Work as a plug-in preprocessing module before image manipulation forensic;
- Deal with manipulation under real senarios with complicated post-processing, achieving robust forencis of quality compression and re-scaling.
Phone Source Identification
- 1:N retrieval and 1:1 alignment via phone-shot-picture;
- Used for image leak tracing, real person authentication, stolen image identification;
- Show robustness to real business data collected online and generalization to unseen mobile devices.