Junao Shen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Junao Shen has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Tian Feng
School of Software Technology, Zhejiang University, Ningbo, China
11 shared publications - Haojie Dong
Zhejiang University
5 shared publications - Jinkang Ji
Zhejiang University
4 shared publications - Weitao Li
School of Software Technology, Zhejiang University, Hangzhou, China
4 shared publications - Bo Li
School of Software Technology, Zhejiang University, Hangzhou, China
3 shared publications - Long Li
School of Software Technology, Zhejiang University, Hangzhou, China
2 shared publications
Latest Publications
- 5 October 2026
- Published by Springer Nature
- p. 404-421
ReInGS: Re-initializing 3D Gaussians Against Sparsity Discrepancy in Few-Shot Novel View Synthesis
- 5 October 2026
- Published by Springer Nature
- p. 404-421
ReInGS: Re-initializing 3D Gaussians Against Sparsity Discrepancy in Few-Shot Novel View Synthesis
- 14 September 2026
- Published by Elsevier in Computers & Graphics
- Vol. 140
ReGAGS: Toward high-fidelity sparse view synthesis via regularizing 3D Gaussians against geometric and appearance overfitting
- 14 September 2026
- Published by Elsevier in Computers & Graphics
- Vol. 140
ReGAGS: Toward high-fidelity sparse view synthesis via regularizing 3D Gaussians against geometric and appearance overfitting
- 4 August 2026
- Published by Elsevier in Computers & Graphics
- Vol. 139
FASS: Frequency-Aware Student Splatting for sparse view synthesis
- 4 August 2026
- Published by Elsevier in Computers & Graphics
- Vol. 139
FASS: Frequency-Aware Student Splatting for sparse view synthesis
- 27 June 2026
- Published by Elsevier in Computers & Graphics
- Vol. 138
CAStGS: Context-aware scene stylization with 3D Gaussian splatting
- 27 June 2026
- Published by Elsevier in Computers & Graphics
- Vol. 138
CAStGS: Context-aware scene stylization with 3D Gaussian splatting
- 11 March 2026
- Published by Elsevier in Computers & Graphics
- Vol. 136
SeCuRe: Toward sparse-view 3D curve reconstruction with Gaussian Splatting
- 11 March 2026
- Published by Elsevier in Computers & Graphics
- Vol. 136
SeCuRe: Toward sparse-view 3D curve reconstruction with Gaussian Splatting
- 4 October 2025
- Published by Elsevier in Computers & Graphics
- Vol. 133
CaRoLS: Condition-adaptive multi-level road layout synthesis
- 4 October 2025
- Published by Elsevier in Computers & Graphics
- Vol. 133
CaRoLS: Condition-adaptive multi-level road layout synthesis
- 5 August 2025
- Published by Elsevier in Computers & Graphics
- Vol. 131
WETR: Wireframe parsing using deformable Transformers
- 5 August 2025
- Published by Elsevier in Computers & Graphics
- Vol. 131
WETR: Wireframe parsing using deformable Transformers
- 5 August 2025
- Published by Elsevier in Computers & Graphics
- Vol. 131
SuraGS: Toward efficient few-shot novel view synthesis via surface-aware Gaussian splatting
- 5 August 2025
- Published by Elsevier in Computers & Graphics
- Vol. 131
SuraGS: Toward efficient few-shot novel view synthesis via surface-aware Gaussian splatting
- Open Access
- 30 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 237-245
DRoLaS: Diffusion-Based Coarse-to-Fine Conditional Synthesis of Hierarchical Road Layouts
- 30 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 237-245
Open AccessDRoLaS: Diffusion-Based Coarse-to-Fine Conditional Synthesis of Hierarchical Road Layouts
- Open Access
- 11 April 2025
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 39 (7) , 6814-6822
In2NeCT: Inter-class and Intra-class Neural Collapse Tuning for Semantic Segmentation of Imbalanced Remote Sensing Images
Remote sensing images (RSIs) are frequently characterized by multi-scale inter-class objects and inconsistently distributed objects due to scene limitations, which would cause a significant data imbalance challenging the corresponding semantic segmentation. Recent methods have leveraged various deep learning techniques to capture high-quality representations for RSI semantic segmentation, but are hardly capable of addressing the afore-mentioned challenge given their limited explorations towards the mechanisms behind the representations. The recently discovered Neural Collapse (NC) phenomenon in computer vision models suggests the simplex equiangular tight frame (ETF) as the optimal representation structure, which has motivated us to observe that the optimal structure of last-layer representations is disrupted and inter-class representations for minor classes tend to become closer to each other beacuse of data imbalance. To address these issues, we propose Inter-class and Intra-class Neural Collapse Tuning (In2NeCT) to optimize the representations that satisfy the simplex ETF, which facilitates the discrimination of inter-class representations and the coherence of intra-class representations. Extensive experiments on three datasets demonstrate that our In2NeCT consistently leads to significant improvements in performance and outperforms the state-of-the-art methods.
- 11 April 2025
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 39 (7) , 6814-6822
Open AccessIn2NeCT: Inter-class and Intra-class Neural Collapse Tuning for Semantic Segmentation of Imbalanced Remote Sensing Images
Remote sensing images (RSIs) are frequently characterized by multi-scale inter-class objects and inconsistently distributed objects due to scene limitations, which would cause a significant data imbalance challenging the corresponding semantic segmentation. Recent methods have leveraged various deep learning techniques to capture high-quality representations for RSI semantic segmentation, but are hardly capable of addressing the afore-mentioned challenge given their limited explorations towards the mechanisms behind the representations. The recently discovered Neural Collapse (NC) phenomenon in computer vision models suggests the simplex equiangular tight frame (ETF) as the optimal representation structure, which has motivated us to observe that the optimal structure of last-layer representations is disrupted and inter-class representations for minor classes tend to become closer to each other beacuse of data imbalance. To address these issues, we propose Inter-class and Intra-class Neural Collapse Tuning (In2NeCT) to optimize the representations that satisfy the simplex ETF, which facilitates the discrimination of inter-class representations and the coherence of intra-class representations. Extensive experiments on three datasets demonstrate that our In2NeCT consistently leads to significant improvements in performance and outperforms the state-of-the-art methods.