Tao Guan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Tao Guan has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Yuesong Wang
Huazhong University of Science and Technology, Wuhan, China
5 shared publications - Luoyuan Xu
Huazhong University of Science and Technology, Wuhan, China
4 shared publications - Yawei Luo
Zhejiang University, Ningbo, China
3 shared publications - Zhuo Chen
Huazhong University of Science and Technology, Wuhan, China
3 shared publications - Bin Zhu
Wuhan Farsee2 Technology Co., Ltd
2 shared publications - Wei Yang
Huazhong University of Science and Technology, Wuhan, China
2 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 5856-5865
D 2 PO: Dual-Direction Preference Optimization via Adaptive Distribution Alignment for Vision-Language Models
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 5856-5865
Open AccessD 2 PO: Dual-Direction Preference Optimization via Adaptive Distribution Alignment for Vision-Language Models
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 4038-4047
MoVT: Video-Augmented Motion Tokenizer for Text-to-Motion Generation
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 4038-4047
Open AccessMoVT: Video-Augmented Motion Tokenizer for Text-to-Motion Generation
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 33 (3) , 109-119
MAStGS: MASt3R-Assisted Efficient 3-D Gaussian Splatting for Indoor Scene Reconstruction
3D Gaussian Splatting (3DGS) achieves impressive rendering quality, speed, and efficiency but struggles in indoor scenes, especially in textureless or weakly textured regions where geometry is hard to capture accurately. We propose MAStGS, a novel method that marries the strengths of MASt3R with 3DGS to enhance geometric reconstruction in indoor environments. MASt3R provides strong geometric priors but is limited by computational costs. To address this, we segment scenes for efficient MASt3R processing, generate a coarse reconstruction, and align it with COLMAP to initialize 3DGS. A coarse-to-fine training strategy further refines geometry using high-resolution depth and normal supervision. Experiments on indoor datasets show that MAStGS achieves more accurate geometry and higher rendering quality than existing 3DGS methods.
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 33 (3) , 109-119
MAStGS: MASt3R-Assisted Efficient 3-D Gaussian Splatting for Indoor Scene Reconstruction
3D Gaussian Splatting (3DGS) achieves impressive rendering quality, speed, and efficiency but struggles in indoor scenes, especially in textureless or weakly textured regions where geometry is hard to capture accurately. We propose MAStGS, a novel method that marries the strengths of MASt3R with 3DGS to enhance geometric reconstruction in indoor environments. MASt3R provides strong geometric priors but is limited by computational costs. To address this, we segment scenes for efficient MASt3R processing, generate a coarse reconstruction, and align it with COLMAP to initialize 3DGS. A coarse-to-fine training strategy further refines geometry using high-resolution depth and normal supervision. Experiments on indoor datasets show that MAStGS achieves more accurate geometry and higher rendering quality than existing 3DGS methods.
- 17 October 2025
- Published by Elsevier in Pattern Recognition
- Vol. 172
Matching ambiguity-resilient multi-view stereo via adaptive patch deformation
- 17 October 2025
- Published by Elsevier in Pattern Recognition
- Vol. 172
Matching ambiguity-resilient multi-view stereo via adaptive patch deformation
- 1 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 32 (4) , 80-90
High-Quality Dynamic Human Novel View Synthesis Based on Multiview Video
- 1 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 32 (4) , 80-90
High-Quality Dynamic Human Novel View Synthesis Based on Multiview Video
- Top Cited Papers
- 20 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 32 (1) , 61-71
EfficientGS: Streamlining Gaussian Splatting for Large-Scale High-Resolution Scene Representation
3D Gaussian Splatting (3DGS) has become a pivotal technology for 3D scene representation. However, its application to large-scale, high-resolution scenes (exceeding 4k×4k pixels) is hindered by the excessive computational and storage requirements for managing a large number of Gaussians. To address these challenges, we propose ’EfficientGS’, an advanced approach that optimizes 3DGS for high-resolution, large-scale scenes. First, we design a more selective densification to reduce the overproliferation of Gaussians. Second, we develop a pruning mechanism to remove redundant Gaussians that only serve as auxiliaries to adjacent ones. Finally, we introduce a sparse order increment strategy for Spherical Harmonics to reduce unnecessary storage. Our empirical evaluations on various datasets, including extensive 4K+ aerial images, demonstrate that ’EfficientGS’ can achieve high rendering fidelity with less computational and storage cost. Our code is available at https://github.com/wenkaier/EfficientGS.
- 20 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 32 (1) , 61-71
Top Cited PapersEfficientGS: Streamlining Gaussian Splatting for Large-Scale High-Resolution Scene Representation
3D Gaussian Splatting (3DGS) has become a pivotal technology for 3D scene representation. However, its application to large-scale, high-resolution scenes (exceeding 4k×4k pixels) is hindered by the excessive computational and storage requirements for managing a large number of Gaussians. To address these challenges, we propose ’EfficientGS’, an advanced approach that optimizes 3DGS for high-resolution, large-scale scenes. First, we design a more selective densification to reduce the overproliferation of Gaussians. Second, we develop a pruning mechanism to remove redundant Gaussians that only serve as auxiliaries to adjacent ones. Finally, we introduce a sparse order increment strategy for Spherical Harmonics to reduce unnecessary storage. Our empirical evaluations on various datasets, including extensive 4K+ aerial images, demonstrate that ’EfficientGS’ can achieve high rendering fidelity with less computational and storage cost. Our code is available at https://github.com/wenkaier/EfficientGS.
- 29 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 30 (1) , 7-16
Learning 3-D Face Shape From Diverse Sources With Cross-Domain Face Synthesis
Monocular face reconstruction is a significant task in many multimedia applications. However, learning-based methods unequivocally suffer from the lack of large datasets annotated with 3D ground-truth. To tackle this problem, we proposed a novel end-to-end 3D face reconstruction network consisting of a domain-transfer conditional GAN (cGAN) and a face reconstruction network. Our method first uses cGAN to translate the realistic face images to the specific rendered style, with a novel 2D facial edge consistency loss function to exploit in-the-wild images. The domain-transferred images are then fed into 3D face reconstruction network. We further propose a novel reprojection consistency loss to restrict 3D face reconstruction network in a self-supervised way. Our approach can be trained with annotated dataset, synthetic dataset and in-the-wild images to learn a unified face model. Extensive experiments have demonstrated the effectiveness of our method.
- 29 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE MultiMedia
- Vol. 30 (1) , 7-16
Learning 3-D Face Shape From Diverse Sources With Cross-Domain Face Synthesis
Monocular face reconstruction is a significant task in many multimedia applications. However, learning-based methods unequivocally suffer from the lack of large datasets annotated with 3D ground-truth. To tackle this problem, we proposed a novel end-to-end 3D face reconstruction network consisting of a domain-transfer conditional GAN (cGAN) and a face reconstruction network. Our method first uses cGAN to translate the realistic face images to the specific rendered style, with a novel 2D facial edge consistency loss function to exploit in-the-wild images. The domain-transferred images are then fed into 3D face reconstruction network. We further propose a novel reprojection consistency loss to restrict 3D face reconstruction network in a self-supervised way. Our approach can be trained with annotated dataset, synthetic dataset and in-the-wild images to learn a unified face model. Extensive experiments have demonstrated the effectiveness of our method.
- 19 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 32 (12) , 8383-8393
Transformer-Based 3D Face Reconstruction With End-to-End Shape-Preserved Domain Transfer
Learning-based face reconstruction methods have recently shown promising performance in recovering face geometry from a single image. However, the lack of training data with 3D annotations severely limits the performance. To tackle this problem, we proposed a novel end-to-end 3D face reconstruction network consisting of a conditional GAN (cGAN) for cross-domain face synthesis and a novel mesh transformer for face reconstruction. Our method first uses cGAN to translate the realistic face images to the specific rendered style, with a 2D facial edge consistency loss function. The domain-transferred images are then fed into face reconstruction network which uses a novel mesh transformer to output 3D mesh vertices. To exploit the domain-transferred in-the-wild images, we further propose a reprojection consistency loss to restrict face reconstruction network in a self-supervised way. Our approach can be trained with annotated dataset, synthetic dataset and in-the-wild images to learn a unified face model. Extensive experiments have demonstrated the effectiveness of our method.
- 19 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 32 (12) , 8383-8393
Transformer-Based 3D Face Reconstruction With End-to-End Shape-Preserved Domain Transfer
Learning-based face reconstruction methods have recently shown promising performance in recovering face geometry from a single image. However, the lack of training data with 3D annotations severely limits the performance. To tackle this problem, we proposed a novel end-to-end 3D face reconstruction network consisting of a conditional GAN (cGAN) for cross-domain face synthesis and a novel mesh transformer for face reconstruction. Our method first uses cGAN to translate the realistic face images to the specific rendered style, with a 2D facial edge consistency loss function. The domain-transferred images are then fed into face reconstruction network which uses a novel mesh transformer to output 3D mesh vertices. To exploit the domain-transferred in-the-wild images, we further propose a reprojection consistency loss to restrict face reconstruction network in a self-supervised way. Our approach can be trained with annotated dataset, synthetic dataset and in-the-wild images to learn a unified face model. Extensive experiments have demonstrated the effectiveness of our method.
- 30 June 2022
- Published by Korean Society for Internet Information (KSII) in KSII Transactions on Internet and Information Systems
- Vol. 16 (6) , 2044-2059
EpiLoc: Deep Camera Localization Under Epipolar Constraint
KSII Transactions on Internet and Information Systems TIIS Paper Details
- 30 June 2022
- Published by Korean Society for Internet Information (KSII) in KSII Transactions on Internet and Information Systems
- Vol. 16 (6) , 2044-2059
EpiLoc: Deep Camera Localization Under Epipolar Constraint
KSII Transactions on Internet and Information Systems TIIS Paper Details