Humphrey Shi
Associate Professor
Works on computer vision and generative models, including image editing and efficient multimodal AI systems.
Humphrey Shi's lab at Georgia Tech works on image & video generation, vision-language & multimodal and multimodal & grounded language. 77 papers since 2023, 48 at top venues, mostly CVPR and WACV. Applications close .
At a glance
- papers, 2023–2026
- 20/ year
- at T1 venues
- 62%
- PhD students
- 0
- MS students
- 0
- alumni listed
- —
Papers by yearfrom DBLP, 2023–2026
- 6 papers2026 · T1 4 · collecting
- 17 papers2025 · T1 13
- 29 papers2024 · T1 16
- 25 papers2023 · T1 15
T1 venuesOtherVenues: CVPR, WACV, ICCV, NeurIPS, ICLR
Research focuspaper titles that use each topic's words
- Image & video generation35 papers
- Vision-language & multimodal11 papers
- Multimodal & grounded language11 papers
- image editing1 paper
- editing1 paper
Paper list
From DBLP.
20266 papers
- WACVBeyond Realism: Learning the Art of Expressive Composition with StickerNet (opens in a new tab)
- CVPR (T1)DuoGen: Towards Autonomous Interleaved Multimodal Generation (opens in a new tab)
- CVPR (T1)MapReduce LoRA: Advancing the Pareto Front in Multi-Preference Optimization for Generative Models (opens in a new tab)
- CVPR (T1)PAI-Bench: A Comprehensive Benchmark For Physical AI (opens in a new tab)
- WACVSafe Vision-Language Models via Unsafe Weights Manipulation (opens in a new tab)
- CVPR (T1)SAGE: Training Smart Any-Horizon Agents for Long Video Reasoning with Reinforcement Learning (opens in a new tab)
202517 papers
- ICLR (T1)ClassDiffusion: More Aligned Personalization Tuning with Explicit Class Guidance (opens in a new tab)
- ICCV (T1)CLIP-GS: Unifying Vision-Language Representation with 3D Gaussian Splatting (opens in a new tab)
- IJCV (T1)Combating Label Noise with a General Surrogate Model for Sample Selection (opens in a new tab)
- CVPR WorkshopDistilling Normalizing Flows (opens in a new tab)
- ICLR (T1)Eagle: Exploring The Design Space for Multimodal LLMs with Mixture of Encoders (opens in a new tab)
- CVPR WorkshopEfficient Image Generation with Variadic Attention Heads (opens in a new tab)
- NeurIPS (T1)Elevating Visual Perception in Multimodal LLMs with Visual Embedding Distillation (opens in a new tab)
- CVPR (T1)Everything to the Synthetic: Diffusion-driven Test-time Adaptation via Synthetic-Domain Alignment (opens in a new tab)
- NeurIPS (T1)FlexVAR: Flexible Visual Autoregressive Modeling without Residual Prediction (opens in a new tab)
- TPAMI (T1)GradBias: Unveiling Word Influence on Bias in Text-to-Image Generative Models (opens in a new tab)
- ICLR (T1)HD-Painter: High-Resolution and Prompt-Faithful Text-Guided Image Inpainting with Diffusion Models (opens in a new tab)
- ICCV (T1)HyPiDecoder: Hybrid Pixel Decoder for Efficient Segmentation and Detection (opens in a new tab)
202429 papers
- ECCV (T1)Benchmarking Object Detectors with COCO: A New Path Forward (opens in a new tab)
- CVPR (T1)Brush2Prompt: Contextual Prompt Generator for Object Inpainting (opens in a new tab)
- ECCV (T1)Collaborative Vision-Text Representation Optimizing for Open-Vocabulary Segmentation (opens in a new tab)
- WACVContinuous Adaptation for Interactive Segmentation Using Teacher-Student Architecture (opens in a new tab)
- NeurIPS (T1)CuMo: Scaling Multimodal LLM with Co-Upcycled Mixture-of-Experts (opens in a new tab)
- CPALDecoding Micromotion in Low-dimensional Latent Spaces from StyleGAN (opens in a new tab)
- CVPR WorkshopDGBD: Depth Guided Branched Diffusion for Comprehensive Controllability in Multi-View Generation (opens in a new tab)
- ECCV (T1)Diffusion for Natural Image Matting (opens in a new tab)
- IEEE Trans. Circuits Syst. Video Technol.FaceCLIP: Facial Image-to-Video Translation via a Brief Text Description (opens in a new tab)
- WACVFarSight: A Physics-Driven Whole-Body Biometric System at Large Distance and Altitude (opens in a new tab)
- NeurIPS (T1)Faster Neighborhood Attention: Reducing the O(n^2) Cost of Self Attention at the Threadblock Level (opens in a new tab)
- NeurIPS (T1)FineStyle: Fine-grained Controllable Style Personalization for Text-to-image Models (opens in a new tab)
202325 papers
- CVPR (T1)Automatic High Resolution Wire Segmentation and Removal (opens in a new tab)
- AAAI (T1)Boosted Dynamic Neural Networks (opens in a new tab)
- TIP (T1)Broad Spectrum Image Deblurring via an Adaptive Super-Network (opens in a new tab)
- TPAMI (T1)CCNet: Criss-Cross Attention for Semantic Segmentation (opens in a new tab)
- TIP (T1)Collaborative Content-Dependent Modeling: A Return to the Roots of Salient Object Detection (opens in a new tab)
- CVPR WorkshopConvMLP: Hierarchical Convolutional MLPs for Vision (opens in a new tab)
- CVPR (T1)Graph Transformer GANs for Graph-Constrained House Generation (opens in a new tab)
- WACVImage Completion with Heterogeneously Filtered Spectral Hints (opens in a new tab)
- AISTInteractive Image Segmentation with Superpixel Propagation (opens in a new tab)
- Comput. Vis. Image Underst.Interactive Neural Painting (opens in a new tab)
- WACVKeys to Better Image Inpainting: Structure and Texture Go Hand in Hand (opens in a new tab)
- NeurIPS (T1)Learning Mask-aware CLIP Representations for Zero-Shot Segmentation (opens in a new tab)
Where alumni went
Alumni are not listed on the lab homepage.
Apply to Georgia TechFall 2027
Georgia Tech Computer Science PhD
- Round
- Single round
- Opens
- —
- GRE
- optional
- English
- TOEFL/IELTS
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNPeking
Zhouchen Lin
- Learning theory & optimization
- CNZhejiang
Jianke Zhu
- Recommendation, search & web
- CNZhejiang
Fei Wu
- Knowledge, QA & retrieval
- USStanford
Maneesh Agrawala
- Rendering & appearance
- SGNTU
Adams Wai-Kin Kong
- Recognition, detection & segmentation
- USUIUC
Alexander Schwing
- Vision-language & multimodal
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.
- The lab site and the paper record do not agree.
- Collected
- Verified
- Papers
- DBLP
- Members
- Lab homepage
- Alumni
- Not collected
Tiers are venue classifications, not lab rankings.
Georgia Techcloses