Research & Publications - Chengjiang Long

Research Interests

Applications in the field of Computer Vision, Computer Graphics, Multimedia, Machine Learning, and Artificial Intelligence
People & Perception AI
- Integration of Human and Machine Intelligence (e.g., collaborative active learning)
- Human Sensing from Unconstrained Images and Videos (e.g., facial expression, motion analysis, and trajectory prediction)
- Object, Scene Recognition and Segmentation
- Camera Pose Estimation and Relocalization
Generative AI
- Generalized Foundation Modeling for Image, Video and 3D Assets Generation, etc.
- Explore Latent Space to Ensure Better Control Signals for Generative Applications (e.g., Disentanglement, ID-Preserving).
- 2D & 3D Digital Human and Related Applications (e.g., Talking Face, Virtual Try-On)
- Photo-Realistic and Cartoon-Style Avatar Generation (e.g., Face, Hair, Head, Hand, and Full-Body Avatars)
- Neural Rendering Related Various Techniques and Applications (e.g., Nerf and Gaussian Splatting)
- Shadow and Illumination Modeling (e.g., Shadow Removal, Delighting, Relighting, and Shadow Generating).
- Multimodal Fusion, Text-to-Image Generation, and Image/Video Caption.
- Style Transfer and Controllable Person Image Synthesis.
- Image and Video Recovery and Enhancement (e.g., Super Resolution).
- Multimodal Large-Scale Language Models (mLLMs), Visual-Language Models (VLMs), Visual-Language-Action Models (VLAs), World Models (WMs), and World Action Models (WAMs) related Technologies and Applications.
Large-Scale Visual and Multimodality Modeling Design, Optimization and Efficiency
- Deep Understanding on SOTA architectures (e.g., Transformer, Diffusion, Autoregression Models).
- Universal Representation Learning and Alignment at Latent Space.
- Cross- and Universal-Domain Adaptation.
- Model Ensembling (e.g., Mixture-of-Models).
- Model Distillation (e.g., Teacher-Student Modeling).
- Low-Rank (e.g., LoRA), Sparsity, and Model Quantization.
- Zero-Shot, Few-Shot, Meta-Learning for Efficient Personalization/Individual Applications.
- Reinforcement Learning, Chain-of-Thoughts, Reasoning, Agentic Workflows/Systems, and Recursive Self-Improvement (RSI) for Various Applications.
Personalized Information Retrieval and Recommendation System
- Multi-Modality Content Understanding
- User Behavior and Engagement Analysis
- Retrieval, Ranking, and Post-Ranking Modeling and Strategies
- Item & User Coldstart Strategies
- Other Recommendation Strategies Related Topics
General AI-Related Topics
- Explainable Artificial Intelligence.
- Retrieval-Augmented Generation (RAG).
- AI Agents and Related Applications.
- AI-Powered Creative Tools and Various Applications (e.g., Search and Chatbot).

Book and Book Chapters

Chengjiang Long, Arslan Basharat, and Anthony Hoogs. "Video Frame Deletion and Duplication", in the book Multimedia Forensics (Chapter 13), edited by Husrev Taha Sencar, Luisa Verdoliva, and Nasir Memon, Springer, 2022.

Selected Publications

The following are selected journal and conference papers from 96 co-author papers, highlighting 1 TOG, 2 T-PAMI, 1 IJCV, 2 TVCG, 2 TIP, 2 T-MM, 1 SIGGRAPH Asia, 13 CVPR, 12 ICCV, 4 ECCV, 6 AAAI, 3 ACM MM, 2 NeurIPS, 1 ICLR, 1 MICCAI, 1 TCSVT, 1 PR, 1 TOMM, 1 CAGD, 3 CGF, 1 RA-L, 2 WACV, 2 ICME, 1 ACCV and 1 ICTAI with 46 1st/2nd-author papers, in which the 2nd authorship indicates the 1st author as my advisors I worked with or as my students/interns/colleagues I closely supervised. For the full list, you can find it in My Google Scholar.

A Training-Free Video Moment Retrieval Framework via Selection from Multiple Visual Prompts

Xun Mo, Shengtao Guo, Yongwei Nie, Fei Ma, Xuemiao Xu, and Chengjiang Long.

Proceedings of the 40-th Annual Conference on Neural Information Processing Systems (NeurIPS), Sydney (Main), Atlanta/Paris (Satellite), Dec 6-13, 2026.

[PDF] [BibTeX]

(Acceptance Rate: 25.7%)

TAR: Temporal Anchor-Constrained Reasoning for Video Temporal Grounding

Chaohong Guo, Xun Mo, Yongwei Nie, Fei Ma, Xuemiao Xu, and Chengjiang Long.

Proceedings of the 19th European Conference on Computer Vision (ECCV), Malmo, Sweden, Sept 8-12, 2026.

[PDF] [BibTeX]

(Acceptance Rate: 27.52%)

RDE-Seg: Role-Disentangled Experts with Residual Routing and Anatomy Constraints for DSA Guidewire Segmentation

Yuning Zhang, Chengjiang Long, Wenju Xu, Tingbao Zhang, Yu Feng, and Jianhui Zhao

Proceedings of 29th International Conference On Medical Image Computing And Computer Assisted Intervention (MICCAI), Starasbourg, France, Sept 27-Oct 1, 2026.

[PDF] [BibTeX]

(Acceptance Rate: 27.51%)

Unlocking Single-View Constraints for Efficient Camera Relocalization with Keypoint-Level Multi-View Geometric Consistency in Training

Hu Lin, Chengjiang Long, Jiqing Zhang, Chuanlu Jiang, Chuanlu_Jiang, Huilin Ge, Erwei Yin, Baocai Yin, and Xin Yang

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Denver, CO, June 3-7, 2026.

[PDF] [BibTeX]

(CVPR 2026 Finding Paper)

T2SGrid: Temporal-to-Spatial Gridification for Video Temporal Grounding

Chaohong Guo, Yihan He, Yongwei Nie, Fei Ma, Xuemiao Xu, and Chengjiang Long

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Denver, CO, June 3-7, 2026.

[PDF] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 25.42%)

FlexAvatar: Flexible Large Reconstruction Model for Animatable Gaussian Head Avatars with Detailed Deformation

Cheng Peng, Zhuo Su, Liao Wang, Chen Guo, Zhaohu Li, Chengjiang Long, Zheng Lv, Jingxiang Sun, Chenyangguang Zhang, and Yebin Liu

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Denver, CO, June 3-7, 2026.

[PDF] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 25.42%)

Invert4TVG: A Temporal Video Grounding Framework with Inversion Tasks for Enhanced Action Understanding

Chenzhaoyu, Hongnan Lin, Yongwei Nie, Fei Ma, Xuemiao Xu, Fei Yu, and Chengjiang Long

The Fourteenth International Conference on Learning Representations (ICLR), April 23-27, 2026.

[PDF] [Poster] [BibTeX]

(Top Conference in Representation Learning, acceptance rate: 28%)

Topology-Guided Semantic Face Center Estimation for Rotation-Invariant Face Detection

Hathai Kaewkorn, Lifang Zhou, Weisheng Li, and Chengjiang Long

IEEE Transactions on Image Processing (TIP), 2026.

[PDF] [BibTeX]

Repurposing 2D Diffusion Models with Gaussian Atlas for 3D Generation

Tiange Xiang, Kai Li, Chengjiang Long, Christian Hane, Peihong Guo, Scott Delp, Ehsan Adeli, and Li Fei-Fei

Proceedings of the International Conference on Computer Vision (ICCV), Oct 19–23, 2025, Honolulu, Hawaii.

[PDF] [Poster] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 24.8%)

DNF-Intrinsic: Deterministic Noise-Free Diffusion for Indoor Inverse Rendering

Rongjia Zheng, Qing Zhang, Chengjiang Long, and Wei-shi Zheng

Proceedings of the International Conference on Computer Vision (ICCV), Oct 19–23, 2025, Honolulu, Hawaii.

[PDF] [Poster] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 24.8%)

PUMPS: Skeleton-Agnostic Point-based Universal Motion Pre-Training for Synthesis in Human Motion Tasks

Clinton Ansun Mo, Kun Hu, Chengjiang Long, and Zhiyong Wang

Proceedings of the International Conference on Computer Vision (ICCV), Oct 19–23, 2025, Honolulu, Hawaii.

[PDF] [Poster] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 24.8%)

PhysSplat: Efficient Physics Simulation for 3D Scenes via MLLM-Guided Gaussian Splatting

Haoyu Zhao, Hao Zhao, Xinyue Zhao, Hao Fei, Hongqui Wang, Chengjiang Long, and Hua Zou

Proceedings of the International Conference on Computer Vision (ICCV), Oct 19–23, 2025, Honolulu, Hawaii.

[PDF] [Poster] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 24.8%)

Incorporating Test-Time Optimization into Training with Dual Networks for Human Mesh Recovery

Yongwei Nie, Mingxian Fan, Chengjiang Long, Qing Zhang, Jian Zhu, and Xuemiao Xu

Proceedings of the thirty-eighth Annual Conference on Neural Information Processing Systems (NeurIPS), Vancouver, Canada, Dec 9-15, 2024.

[PDF] [Poster] [BibTeX]

(Top Conference in Machine Learning, acceptance rate: 25.8%)

Exploring Matching Rates: From Key Point Selection to Camera Relocalization

Hu Lin, Chengjiang Long, Yifeng Fei, qianchen xia, Erwei Yin, Baocai Yin, and Xin Yang

Proceedings of the 32nd ACM Multimedia Conference (ACM MM), Melbourne, Australia, Oct 28-Nov 1, 2024.

[PDF][BibTeX]

(Top Conference in Multimedia, acceptance rate: 3.97%, Oral Paper!)

Motion Keyframe Interpolation for Any Human Skeleton using Point Cloud-based Human Motion Data Homogenisation

Clinton Mo, Kun Hu, Chengjiang Long, Dong Yuan, and Zhiyong Wang

Proceedings of the European Conference on Computer Vision (ECCV), MiCo Milano, Milan, Italy, Sep 28-Oct 4, 2024.

[PDF] [Supp] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 18.95%)

Multi-RoI Human Mesh Recovery with Camera Consistency and Contrastive Losses

Yongwei Nie, Changzhen Liu, Chengjiang Long, Qing Zhang, Guiqing Li, and Hongmin Cai

Proceedings of the European Conference on Computer Vision (ECCV), MiCo Milano, Milan, Italy, Sep 28-Oct 4, 2024.

[PDF] [Supp] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 18.95%)

Interleaving One-Class and Weakly-Supervised Models with Adaptive Thresholding for Unsupervised Video Anomaly Detection

Yongwei Nie, Hao Huang, Chengjiang Long, Qing Zhang, Pradipta Maji, and Hongmin Cai

Proceedings of the European Conference on Computer Vision (ECCV), MiCo Milano, Milan, Italy, Sep 28-Oct 4, 2024.

[PDF] [Supp] [BibTeX]

(Top 3 Conference in Computer Vision, acceptance rate: 18.95%)

Single-Image SVBRDF Estimation Using Auxiliary Renderings as Intermediate Targets

Yongwei Nie, Jiaqi Yu, Chengjiang Long, Qing Zhang, Guiqing Li, and Hongmin Cai

IEEE Transactions on Visualization and Computer Graphics (TVCG), 2024.

[PDF] [BibTeX]

(Top 2 journal in Computer Graphics, IF: 5.226)

REFA: Real-time Egocentric Facial Animations for Virtual Reality

Chengjiang Long as a part of Meta VR Face Tracking Team

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshop and Competition on Affective Behavior Analysis in-the-wild (CVPR-W), Seattle, WA, Jun 18, 2024.

[PDF] [Supp] [BibTeX]

CoreRec: A Counterfactual Correlation Inference for Next Set Recommendation

Kexin Li, Chengjiang Long, Shengyu Zhang, Xudong Tang, Zhichao Zhai, Kun Kuang, and Jun Xiao

Proceedings of the Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI), Vancouver, BC, Canada, Feb 20-27, 2024.

[PDF] [BibTeX]

(Top 1 Conference in Artificial Intelligence, acceptance rate: 23.75%)

Disentangled Representation Learning for Controllable Person Image Generation

Wenju Xu, Chengjiang Long, Yongwei Nie, and Guanghui Wang

IEEE Transactions on Multimedia (T-MM), 2024.

[PDF] [BibTeX]

(Top 1 Journal in Multimedia, IF: 6.051)

Representing Multimodal Behaviors with Mean Location for Pedestrian Trajectory Prediction

Liushuai Shi, Le Wang, Chengjiang Long, Sanping Zhou, Wei Tang, Nanning Zheng, and Gang Hua

IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2023.

Research Interests

Book and Book Chapters

Selected Publications

A Training-Free Video Moment Retrieval Framework via Selection from Multiple Visual Prompts

TAR: Temporal Anchor-Constrained Reasoning for Video Temporal Grounding

RDE-Seg: Role-Disentangled Experts with Residual Routing and Anatomy Constraints for DSA Guidewire Segmentation

Unlocking Single-View Constraints for Efficient Camera Relocalization with Keypoint-Level Multi-View Geometric Consistency in Training

T2SGrid: Temporal-to-Spatial Gridification for Video Temporal Grounding

FlexAvatar: Flexible Large Reconstruction Model for Animatable Gaussian Head Avatars with Detailed Deformation

Invert4TVG: A Temporal Video Grounding Framework with Inversion Tasks for Enhanced Action Understanding

Topology-Guided Semantic Face Center Estimation for Rotation-Invariant Face Detection

Repurposing 2D Diffusion Models with Gaussian Atlas for 3D Generation

DNF-Intrinsic: Deterministic Noise-Free Diffusion for Indoor Inverse Rendering

PUMPS: Skeleton-Agnostic Point-based Universal Motion Pre-Training for Synthesis in Human Motion Tasks

PhysSplat: Efficient Physics Simulation for 3D Scenes via MLLM-Guided Gaussian Splatting

Incorporating Test-Time Optimization into Training with Dual Networks for Human Mesh Recovery

Exploring Matching Rates: From Key Point Selection to Camera Relocalization

Motion Keyframe Interpolation for Any Human Skeleton using Point Cloud-based Human Motion Data Homogenisation

Multi-RoI Human Mesh Recovery with Camera Consistency and Contrastive Losses

Interleaving One-Class and Weakly-Supervised Models with Adaptive Thresholding for Unsupervised Video Anomaly Detection

Single-Image SVBRDF Estimation Using Auxiliary Renderings as Intermediate Targets

REFA: Real-time Egocentric Facial Animations for Virtual Reality

CoreRec: A Counterfactual Correlation Inference for Next Set Recommendation

Disentangled Representation Learning for Controllable Person Image Generation

Representing Multimodal Behaviors with Mean Location for Pedestrian Trajectory Prediction

Learning Dynamic Style Kernels for Artistic Style Transfer

Feature Representation Learning with Adaptive Displacement Generation and Transformer Fusion for Micro-Expression Recognition

Continuous Intermediate Token Learning with Implicit Motion Manifold for Keyframe based Motion Interpolation

Exploiting Residual and Illumination with GANs for Shadow Detection and Shadow Removal

Explore Contextual Information for 3D Scene Graph Generation

Diverse Human Motion Prediction via Gumbel-Softmax Sampling from an Auxiliary Space

Deep Image-based Illumination Harmonization

Progressively Generating Better Initial Guesses Towards Next Stages for High-Quality Human Motion Prediction

Video Shadow Detection via Spatio-Temporal Interpolation Consistency Training

Social Interpretable Tree for Pedestrian Trajectory Prediction

Complementary Attention Gated Network for Pedestrian Trajectory Prediction

CPRAL: Collaborative Panoptic-Regional Active Learning for Semantic Segmentation

DDBN: Dual Detection Branch Network for Semantic Diversity Predictions

Monte Carlo Denoising via Auxiliary Feature Guided Self-Attention

DRB-GAN: A Dynamic ResBlock Generative Adversarial Network for Artistic Style Transfer

CANet: A Context-Aware Network for Shadow Removal

A Hybrid Video Anomaly Detection Framework via Memory-Augmented Flow Reconstruction and Flow-Guided Frame Prediction

MSR-GCN: Multi-Scale Residual Graph Convolution Networks for Human Motion Prediction

Dual Graph Convolutional Networks with Transformer and Curriculum Learning for Image Captioning

SGCN: Sparse Graph Convolution for Pedestrian Trajectory Prediction

A Novel Visual Representation on Text Using Diverse Conditional GAN for Visual Recognition

A Hybrid Attention Mechanism for Weakly-Supervised Temporal Action Localization

A Two-Stage Attentive Network for Single Image Super Resolution

Luminance Attentive Networks for HDR Image and Panorama Reconstruction

CRD-CGAN: Category-Consistent and Relativistic Constraints for Diverse Text-to-Image Generation

Explore Video Clip Order with Self-Supervised and Curriculum Learning for Video Applications

DOA-GAN: Dual-Order Attentive Generative Adversarial Network for Image Copy-move Forgery Detection and Localization

ARShadowGAN: Shadow Generative Adversarial Network for Augmented Reality in Single Light Scenes

RIS-GAN: Explore Residual and Illumination with Generative Adversarial Networks for Shadow Removal

Multi-Stage Point Completion Network with Critical Set Supervision

CLA-GAN: A Context and Lightness Aware Generative Adversarial Network for Shadow Removal

Multi-Context and Enhanced Reconstruction Network for Single Image Super Resolution

Iterative and Adaptive Sampling with Spatial Attention for Black-Box Model Explanations

ARGAN: Attentive Recurrent Generative Adversarial Network for Shadow Detection and Removal

Shadow Inpainting and Removal Using Generative Adversarial Networks with Slice Convolutions

Automated Pig Counting Using Deep Learning

Deep Neural Networks In Fully Connected CRF For Image Labeling With Social Network Metadata

A Coarse-to-fine Deep Convolutional Neural Network Framework for Frame Duplication Detection and Localization in Forged Videos

Urban Semantic 3D Reconstruction from Multiview Satellite Imagery

VITAL: A Visual Interpretation on Text with Adversarial Learning for Image Labeling

Collaborative Active Visual Recognition from Crowds: A Distributed Ensemble Approach

How Does A Camera Look At One 3D CAD Object?

Correlational Gaussian Processes for Cross-domain Visual Recognition

A C3D-based Convolutional Neural Network for Frame Dropping Detection in a Single Video Shot

L2GSCI: Local to Global Seam Cutting and Integrating for Accurate Face Contour Extraction

A Joint Gaussian Process Model for Active Visual Recognition with Expertise Estimation in Crowdsourcing.

Multi-class Multi-annotator Active Learning with Robust Gaussian Process for Visual Recognition

Accurate Object Detection with Location Relaxation and Regionlets Re-localization

Active Visual Recognition with Expertise Estimation in Crowdsourcing

Collaborative Active Learning of a Kernel Machine Ensemble for Recognition

A New K Nearest Neighbors Search Algorithm Using Cell Grids for 3D Scattered Point Cloud

A New Region Growing Algorithm for Triangular Mesh Recovery from Scattered 3D Points

Transmission: A New Feature for Computer Vision Based Smoke Detection

Improvements on IPD Algorithm for Triangular Mesh Reconstruction from 3D Point Cloud

A New K-Nearest Neighbors Search Algorithm Based on 3D Cell Grids