TY - GEN
T1 - IMPROVING NOVEL VIEW SYNTHESIS OF 360◦ SCENES IN EXTREMELY SPARSE VIEWS BY JOINTLY TRAINING HEMISPHERE SAMPLED SYNTHETIC IMAGES
AU - Chen, Guangan
AU - Truong, Anh Minh
AU - Lin, Hanhe
AU - Vlaminck, Michiel
AU - Philips, Wilfried
AU - Luong, Hiep
N1 - Publisher Copyright:
©2025 IEEE.
PY - 2025/8/18
Y1 - 2025/8/18
N2 - Novel view synthesis in 360◦ scenes from extremely sparse input views is essential for applications like virtual reality and augmented reality. This paper presents a novel framework for novel view synthesis in extremely sparse-view cases. As typical structure-from-motion methods are unable to estimate camera poses in extremely sparse-view cases, we apply DUSt3R to estimate camera poses and generate a dense point cloud. Using the poses of estimated cameras, we densely sample additional views from the upper hemisphere space of the scenes, from which we render synthetic images together with the point cloud. Training 3D Gaussian Splatting model on a combination of reference images from sparse views and densely sampled synthetic images allows a larger scene coverage in 3D space, addressing the overfitting challenge due to the limited input in sparse-view cases. Retraining a diffusion-based image enhancement model on our created dataset, we further improve the quality of the point-cloud-rendered images by removing artifacts. We compare our framework with benchmark methods in cases of only four input views, demonstrating significant improvement in novel view synthesis under extremely sparse-view conditions for 360◦ scenes.
AB - Novel view synthesis in 360◦ scenes from extremely sparse input views is essential for applications like virtual reality and augmented reality. This paper presents a novel framework for novel view synthesis in extremely sparse-view cases. As typical structure-from-motion methods are unable to estimate camera poses in extremely sparse-view cases, we apply DUSt3R to estimate camera poses and generate a dense point cloud. Using the poses of estimated cameras, we densely sample additional views from the upper hemisphere space of the scenes, from which we render synthetic images together with the point cloud. Training 3D Gaussian Splatting model on a combination of reference images from sparse views and densely sampled synthetic images allows a larger scene coverage in 3D space, addressing the overfitting challenge due to the limited input in sparse-view cases. Retraining a diffusion-based image enhancement model on our created dataset, we further improve the quality of the point-cloud-rendered images by removing artifacts. We compare our framework with benchmark methods in cases of only four input views, demonstrating significant improvement in novel view synthesis under extremely sparse-view conditions for 360◦ scenes.
KW - 360 scenes
KW - 3D Gaussian Splatting
KW - diffusion model
KW - extremely sparse views
KW - image enhancement
UR - https://www.scopus.com/pages/publications/105028629348
U2 - 10.1109/ICIP55913.2025.11084363
DO - 10.1109/ICIP55913.2025.11084363
M3 - Conference contribution
AN - SCOPUS:105028629348
SN - 9798331523800
T3 - Proceedings - International Conference on Image Processing, ICIP
SP - 821
EP - 826
BT - 2025 IEEE International Conference on Image Processing, ICIP 2025 - Proceedings
PB - IEEE
CY - USA
T2 - 32nd IEEE International Conference on Image Processing, ICIP 2025
Y2 - 14 September 2025 through 17 September 2025
ER -