Editorial Overview: Shape Modeling International 2023
The editorial for the Computers & Graphics special section on Shape Modeling International 2023 (SMI2023) marks a significant milestone for the geometry processing community. As the first fully in-person gathering following the disruptions of the COVID-19 pandemic, the conference, held in Genoa, Italy, showcased the resilience and rapid evolution of the field. This special section presents 18 high-quality papers that bridge classical mathematical foundations with cutting-edge computational challenges.
Rigorous Selection and the Open Science Initiative
The selection process for SMI2023 was exceptionally stringent to ensure the highest academic standards. Out of 61 full-paper submissions, only 18 were chosen for this journal special section—an acceptance rate of approximately 29.5%. Each paper underwent a two-stage double-blind review process, involving an average of four expert reviewers per submission. The final selection represents a global effort, with contributions from research hubs in China, France, Italy, and the United States.
A standout feature of this year's cycle was the proactive emphasis on Open Science. The editors collaborated closely with the Graphics Replicability Stamp Initiative to encourage authors to share their code and data.
Six of the accepted papers were awarded the "Replicability Badge," signaling a community-wide shift toward transparent and reproducible research. Notably, the organizers prioritized papers with this stamp when considering candidates for the "Best Paper" award, incentivizing a culture of transparency.
Key Research Frontiers: Neural Representations and Digital Twins
The research presented in this section highlights a pivotal shift in shape modeling. While traditional geometry processing—such as mesh optimization and surface analysis—remains a cornerstone, the field is increasingly embracing deep geometric learning.
- The Rise of Neural Representations: The "Best Paper" award was granted to Neural skeleton: implicit neural representation away from the surface by Mattéo Clémot and Julie Digne. Their work utilizes implicit neural representations to address classical problems like skeletonization and surface reconstruction, demonstrating how AI can enhance geometric robustness.
- Digital Twins as a Frontier: The conference explored the frontier of 3D Digital Twins. Keynote discussions on spatio-semantic modeling for urban environments suggest that the community is moving toward large-scale, semantically rich simulations that integrate geometry with real-world data for industrial and urban applications.
Distinguished Awards and Recognition
SMI2023 served as a platform to honor individual excellence and long-standing service within the community:
- Best Paper Award: Neural skeleton: implicit neural representation away from the surface (Clémot & Digne).
- Honorable Mentions: Inpainting Holes in Folded Fabric Meshes (Gisbert et al.) and Discrete Orthogonal Structures (Dellinger et al.).
- Tosiyasu L. Kunii Distinguished Researcher Award: Presented to Wenping Huang for his significant contributions to the field.
- SMI Young Investigator Award: Awarded to Etienne Corman for promising early-career research.
- Alexander Pasko Service Award: The inaugural award was given to Michela Spagnuolo for her long-standing dedication and service to the shape modeling community.
Conclusion and Future Outlook
The SMI2023 special section reflects a field in transition, successfully integrating artificial intelligence with rigorous geometric constraints. By fostering a culture of replicability and expanding into massive-scale digital twins, the shape modeling community continues to define the future of digital representation. Future work is expected to focus on the seamless integration of AI with geometric constraints and the scaling of these techniques to handle the complexity of urban-scale environments.
Summary updated