Jianfu Zhang
Tenured Track Associate Professor
Works on visual content generation and trustworthy AI models.
Jianfu Zhang's lab at SJTU works on image & video generation, trustworthy & safe ML and interpretability, evaluation & safety of LMs. 39 papers since 2023, 24 at top venues, mostly ICASSP and AAAI. 4 PhD students and 2 master's students. Applications close .
At a glance
- papers, 2023–2026
- 10/ year
- at T1 venues
- 62%
- PhD students
- 4
- MS students
- 2
- alumni listed
- —
Papers by yearfrom the lab homepage, 2023–2026
- 15 papers2026 · T1 10 · collecting
- 8 papers2025 · T1 3
- 9 papers2024 · T1 7
- 7 papers2023 · T1 4
T1 venuesOtherVenues: ICASSP, AAAI, ACM MM, CVPR, ECCV
Research focuspaper titles that use each topic's words
- Image & video generation15 papers
- Interpretability, evaluation & safety of LMs1 paper
- content1 paper
- Trustworthy & safe ML0 papers
- content generation0 papers
Paper list
From the lab homepage.
202615 papers
- CVPR (T1)AVGGT: Rethinking Global Attention for Accelerating VGGT
- ICIPCOCO-Inpaint: A Benchmark for Image Inpainting Detection and Manipulation Localization (opens in a new tab)
- ACM MMDefake-o3: From Speculative Rationales to Verifiable Evidence for Explainable AIGI Detection
- Science China Information SciencesDeMamba: AI-Generated Video Detection on Million-Scale GenVideo Benchmark (opens in a new tab)
- ICLR (T1)FakeXplain: AI-Generated Image Detection via Human-Aligned Grounded Reasoning
- ICASSP (T1)FW-VTON: Flattening-and-Warping for Person-to-Person Virtual Try-On (opens in a new tab)
- ICASSP (T1)GAMMA: Generalizable AI-Generated Image Detection via Multi-Task and Manipulation-Augmented Supervision (opens in a new tab)
- AAAI (T1)High-Quality Full-Head 3D Avatar Generation from Any Single Portrait Image (opens in a new tab)
- CVPR (T1)Locate-Then-Examine: Grounded Region Reasoning Improves Detection of AI-Generated Images
- ECCV (T1)LoRC: Detecting AI-Generated Images via Low-Rank Collapse in Semantic Residuals (opens in a new tab)
- ICASSP (T1)PHomo: Patch Homogeneity for No-Reference Inpainting Assessment and Verification (opens in a new tab)
- EMNLP (T1)RoCo-ACE: Rollout-Conditioned Online Distillation for Retention-Aware Knowledge Injection
20258 papers
- Pattern RecognitionDefending adversarial attacks in Graph Neural Networks via tensor enhancement (opens in a new tab)
- NeurIPS (T1)GeneMAN: Generalizable Single-Image 3D Human Reconstruction from Multi-Source Human Data
- CSCWDHyperspectral Image Classification Based on Local Low-Rank Representation (opens in a new tab)
- PRCVID-MotionNet: Identity-Preserved 3D Skeleton Sequence Generation via Information Bottleneck Disentanglement (opens in a new tab)
- ACM MMInterAnimate: Taming Region-Aware Diffusion Model for Realistic Human Interaction Animation (opens in a new tab)
- ICLR (T1)Pedestrian Motion Reconstruction: A Large-scale Benchmark via Mixed Reality Rendering with Multiple Perspectives and Modalities
- ACM MMTowards Explainable Fake Image Detection with Multi-Modal Large Language Models (opens in a new tab)
- AAAI (T1)WildFake: A Large-Scale and Hierarchical Dataset for AI-Generated Images Detection (opens in a new tab)
20249 papers
- SIGIR (T1)An E-Commerce Dataset Revealing Variations during Sales (opens in a new tab)
- ICASSP (T1)Arbitrary Style Transfer with Prototype-Based Channel Alignment (opens in a new tab)
- CIKM (T1)Assessing Image Inpainting via Re-Inpainting Self-Consistency Evaluation (opens in a new tab)
- ECCV (T1)ComFusion: Enhancing Personalized Generation by Instance-Scene Compositing and Fusion (opens in a new tab)
- NeurIPS (T1)DomainGallery: Few-shot Domain-driven Image Generation by Attribute-centric Finetuning
- ICASSP (T1)Hierarchical Attacks on Large-Scale Graph Neural Networks (opens in a new tab)
- Neural Processing LettersMulti-Model UNet: An Adversarial Defense Mechanism for Robust Visual Tracking (opens in a new tab)
- ICASSP (T1)ProAug: Prototype-Based Augmentation for Long-Tailed Image Classification (opens in a new tab)
- CVPR WorkshopSupervised Contrastive Learning for Snapshot Spectral Imaging Face Anti-Spoofing (opens in a new tab)
20237 papers
- AAAI (T1)Amodal Instance Segmentation via Prior-Guided Expansion (opens in a new tab)
- AAAI (T1)Critical Firms Prediction for Stemming Contagion Risk in Networked-Loans through Graph-Based Deep Reinforcement Learning (opens in a new tab)
- Pattern RecognitionDiverse image inpainting with disentangled uncertainty (opens in a new tab)
- AAAI (T1)Isometric Manifold Learning Using Hierarchical Flow (opens in a new tab)
- AAAI (T1)Memorization Weights for Instance Reweighting in Adversarial Training (opens in a new tab)
- Journal of Computer Science and TechnologyNatural Image Matting with Attended Global Context (opens in a new tab)
- ACM MMTaming the Power of Diffusion Models for High-Quality Virtual Try-On with Appearance Flow (opens in a new tab)
Where alumni went
Alumni are not listed on the lab homepage.
Apply to SJTUFall 2027
Computer Science PhD (international applicants)
- Round
- Self-funded
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- USStanford
Maneesh Agrawala
- Rendering & appearance
- GBUCL
Fabio Pierazzi
- Trustworthy & safe ML
- CNPeking
Wen Gao
- Video understanding
- USUSC
C.-C. Jay Kuo
- Video understanding
- USGeorgia Tech
Humphrey Shi
- Image & video generation
- USUMD
Tom Goldstein
- Trustworthy & safe ML
Your notes
Save this lab to keep notes.
Data
Some numbers could not be verified:
- The alumni list could not be found on the lab site.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Not collected
Tiers are venue classifications, not lab rankings.
SJTUcloses