Research Projects

3D Visual Data Coding

The research work includes 3D point cloud coding, gaussian splatting compression and rendering, image/video coding, multimodal coding, quality assessment and enhancement, and the related standardization efforts. We mainly focus on devising advanced deep learning solutions for coding optimization, and various methods to obtain better rate-distortion performance with flexibility and scalability. We have also conducted research works including light-weight algorithms and models, hardware acceleration and implementations, and open source projects. We have recently published high quality papers on IEEE TPAMI, IEEE TIP, IEEE TCSVT, IEEE TMM, IEEE TNNLS, IEEE TGRS, IJCV, CVPR, ICCV, ECCV, ACM MM, AAAI, DCC, etc., and participated into the standardization work of MPEG, AVS, and IEEE. We have also established several open source projects for multimedia computing and AI, including OpenAICoding, OpenPointCloud, OpenDatasets, etc. The monographs titled “AI-based 3D Point Cloud Coding: Methods, Standards, and Applications”, “AI-based Image and Video Coding: Methods, Standards, and Applications”, “Point Cloud Compression: Technologies and Standardization” have been published by Springer Nature in 2026, 2025, and 2024, respectively.

UniPCGC: Towards Practical Point Cloud Geometry Compression via An Efficient Unified Approach
AAAI, 2025.
AdaDPCC: Adaptive Rate Control and Rate-Distortion-Complexity Optimization for Dynamic Point Cloud Compression
AAAI, 2025.
Saliency Segmentation Oriented Deep Image Compression with Novel Bit Allocation
IEEE TIP, 2025.
Stochasticity-aware No-Reference Point Cloud Quality Assessment
IJCAI, 2025.
SPCGC: Scalable Point Cloud Geometry Compression for Machine Vision
ICRA, 2024.
ROI-Guided Point Cloud Geometry Compression Towards Human and Machine Vision
ACM MM, 2024.
Zoom to Perceive Better: No-reference Point Cloud Quality Assessment via Exploring Effective Multiscale Feature
IEEE TCSVT, 2024.
Learned Rate Control for Frame-Level Adaptive Neural Video Compression via Dynamic Neural Network
ECCV, 2024.
AdaNIC: Towards Practical Neural Image and Compression via Dynamic Transform Routing
ICCV, 2024.
Low Complexity Coding Unit Decision in Video-Based Point Cloud Compression
IEEE TIP, 2024.
OpenHardwareVC: An Open Source Library for 8K UHD Video Coding Hardware Implementation
ACM MM, 2024.
OpenDIC: An Open-Source Library and Performance Evaluation for Deep-learning-based Image Compression
ACM MM, 2024.
OpenDMC: An Open-Source Library and Performance Evaluation for Deep-learning-based Multi-frame Compression
ACM MM, 2024.
OpenPointCloud: An Open-Source Algorithm Library of Deep Learning Based Point Cloud Compression
ACM MM, 2024.
Efficient Neural Network Compression Inspired by Compressive Sensing
IEEE TNNLS, 2024.
Interpretable Task-inspired Adaptive Filter Pruning For Neural Networks Under Multiple Constraints
IJCV, 2024.
3D Point Cloud Attribute Compression Using Diffusion-based Texture-aware Intra Prediction
IEEE TCSVT, 2024.
Enlarged Motion-Aware and Frequency-Aware Network for Compressed Video Artifact Reduction
IEEE TCSVT, 2024.
Fast Inter-Frame Motion Prediction for Compressed Dynamic Point Cloud Attribute Enhancement
AAAI, 2024.
End-to-End RGB-D Image Compression via Exploiting Channel-Modality Redundancy
AAAI, 2024.
ViewPCGC: View-Guided Learned Point Cloud Geometry Compression
ACM MM, 2024.
Rate-Distortion-Guided Learning Approach with Cross-Projection Information for V-PCC Fast CU Decision
ACM MM, 2023.
OpenFastVC: An Open Source Library for Video Coding Fast Algorithm Implementation
ACM MM, 2023.
Block-Adaptive Point Cloud Attribute Coding with Region-Aware Optimized Transform
IEEE TCSVT, 2023.
Semantic Point Cloud Upsampling
IEEE TMM, 2023.
SUR-Driven Video Coding Rate Control for Jointly Optimizing Perceptual Quality and Buffer Control
IEEE TIP, 2023.
OctAttention: Octree-based Large-scale Contexts Model for Point Cloud Compression
AAAI, 2022.

3D Visual Data Processing

The research work includes 3D reconstruction and generation, 3D perception and understanding, generative AI and world models, vision-language-action (VLA) models, 3D vision and multimodal learning for spatial intelligence and embodied Intelligence. We have recently published high quality papers on IEEE TPAMI, IEEE TIP, IEEE TCSVT, IEEE TMM, IEEE TNNLS, NeurIPS, ICLR, CVPR, ICCV, ECCV, ACM MM, AAAI, ICRA, etc. The monograph titled “Deep Learning for 3D Point Clouds” has been published by Springer Nature in 2025.

Point Cloud Semantic Segmentation With Sparse and Inhomogeneous Annotations
AAAI, 2025.
VE-Bench: Subjective-Aligned Benchmark Suite for Text-Driven Video Editing Quality Assessment
AAAI, 2025.
Less is More: Label Recommendation for Weakly Supervised Point Cloud Semantic Segmentation
AAAI, 2024.
Point-MPP: Point Cloud Self-supervised Learning from Masked Position Prediction
IEEE TNNLS, 2024.
Improving Graph Representation for Point Cloud Segmentation via Attentive Filtering
CVPR, 2023.
A Thorough Benchmark and A New Model for Light Field Saliency Detection
IEEE TPAMI, 2023.
Unified Information Fusion Network for Multi-Modal RGB-D and RGB-T Salient Object Detection
IEEE TCSVT, 2022.
Salient Object Detection for Point Clouds
ECCV, 2022.