EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

IA-3DGS Identity-Anchored Dynamic Gaussian Splatting for Long-Term Facial Detail Preservation

Ze Zhao iD, Lili Yin iD, Shuaijie Wang iD, Jie Gao

Sensors · 05.08.2026 · 26 (15) · 4947

DOI: 10.3390/s26154947 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 05.08.2026. Verfasst von Ze Zhao und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer Science and Technology, Harbin University of Science and Technology, Harbin 150080, China. Die Arbeit nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Dynamic 3D Gaussian Splatting (3DGS) enables efficient facial-avatar rendering but may suffer from cumulative geometric drift and degradation of identity-specific details when applied to long monocular sequences. To address this problem, we propose IA-3DGS, an identity-anchored dynamic Gaussian framework for long-term facial animation. The proposed framework contains three main components. First, a 3D morphable model-guided semantic initialization strategy associates Gaussian primitives with anatomically meaningful facial regions. Second, a region-weighted Jacobian rigidity regularizer suppresses non-physical shear and anisotropic stretching in quasi-rigid facial regions while preserving sufficient flexibility in expression-sensitive regions. Third, a cyclic memory correction mechanism periodically aligns the current identity representation with a frozen reference representation and applies exponential moving average smoothing to reduce correction-induced temporal discontinuities. Experiments were conducted on the NeRSemble and NHA datasets and on a self-collected Custom-5K dataset containing continuous monocular facial sequences longer than 5000 frames. IA-3DGS achieved a PSNR of 31.85 dB, an SSIM of 0.942, and an LPIPS of 0.038 on standard-length sequences. At frame 5000, the method obtained an L-LPIPS of 0.046 and a landmark mean error of 1.3 mm, compared with 0.245 and 6.8 mm, respectively, for the purely implicit dynamic 3DGS baseline. The system rendered 1920 × 1080 images at approximately 48 frames per second on a single NVIDIA RTX 4090. These results indicate that the proposed semantic, geometric, and temporal constraints improve long-sequence identity consistency under the evaluated subject-specific monocular setting.

Schlagwörter

Stichwörter der Autor:innen

3D Gaussian splattingJacobian regularizationdynamic head avataridentity preservationlong-sequence facial animationmonocular video

Autor:innen und Institutionen

  • School of Computer Science and Technology, Harbin University of Science and Technology, Harbin 150080, China.

ORCID-Profile: Ze Zhao, Lili Yin, Shuaijie Wang

Zitieren

APA

Zhao, Z., Yin, L., Wang, S., & Gao, J. (2026). IA-3DGS Identity-Anchored Dynamic Gaussian Splatting for Long-Term Facial Detail Preservation. Sensors, 26(15), 4947. https://doi.org/10.3390/s26154947

Vancouver

Zhao Z, Yin L, Wang S, Gao J. IA-3DGS Identity-Anchored Dynamic Gaussian Splatting for Long-Term Facial Detail Preservation. Sensors. 2026;26(15):4947. doi:10.3390/s26154947

BibTeX
@article{Zhao2026,
  author = {Zhao, Ze and Yin, Lili and Wang, Shuaijie and Gao, Jie},
  title = {IA-3DGS Identity-Anchored Dynamic Gaussian Splatting for Long-Term Facial Detail Preservation},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {15},
  pages = {4947},
  doi = {10.3390/s26154947}
}
Kostenlos starten →