Message from the 3DV 2020 Program Chairs

Adrian Hilton, Zuzana Kukelova, Stephen Lin, Jun Sato3DV 2020 · DOI

Abstract

The 2020 International Conference on 3D Vision (3DV) marked a significant milestone in the field's evolution, characterized by a massive 62% surge in submissions and a forced transition to a virtual format due to the COVID-19 pandemic. This report summarizes the organizational strategies, peer-review methodology, and program outcomes detailed by the 3DV 2020 Program Chairs. Despite logistical challenges, the conference maintained a rigorous selection process, accepting 123 papers from a record pool of 272 submissions, reflecting the explosive growth of 3D computer vision driven by deep learning and robotics.

Introduction

The International Conference on 3D Vision has long served as a premier venue for research in 3D sensing, modeling, and analysis. In 2020, the conference faced a dual challenge: managing an unprecedented increase in interest from the global research community while simultaneously navigating the disruptions caused by the COVID-19 pandemic. Originally planned for Fukuoka, Japan, the event transitioned to a fully virtual format, necessitating a rethink of how research is disseminated and how community engagement is fostered in a digital environment.

Scaling the Review Process

With 272 valid submissions—a 62% increase over the previous year—the Program Chairs implemented a multi-tiered organizational strategy to ensure high-quality, unbiased peer reviews. The process was designed around several core principles:

  • Impartiality and Conflict Management: To maintain the highest ethical standards, the Program Chairs and the General Chair recused themselves from submitting any papers to the conference.
  • Hierarchical Structure: The review process was managed by 30 Area Chairs (ACs), selected for their diverse expertise, seniority, and geographical representation. These ACs oversaw a pool of 224 reviewers.
  • Workload Management: To prevent reviewer burnout and ensure each paper received a thorough critique, the number of assignments per reviewer was capped at five. A dedicated pool of "emergency reviewers" was also established to handle last-minute needs.
  • Interactive Evaluation: The process included a bidding phase for reviewers to align papers with their expertise, followed by an author rebuttal phase and internal discussions between reviewers and ACs before final recommendations were made.

The Virtual Pivot

The transition from a physical meeting at Kyushu University to a virtual platform required new methods for research presentation. To balance the need for quick information gathering with deep technical dives, the organizers required authors of all accepted papers to provide two distinct video formats:

  1. Short Presentations (3 minutes): Designed for quick "skimming" and high-level conceptual overviews.
  2. Long Presentations (10 minutes): Intended for detailed technical explanations and in-depth methodology discussions.

Key Insights and Program Results

The chairs attributed the surge in submissions to the "Deep Learning Effect"—the rapid expansion of the field fueled by advances in neural networks, 3D sensing technology, and the maturity of application domains like autonomous systems, robotics, and augmented reality (AR).

Program Statistics

The final technical program reflected a competitive selection process:

  • Total Submissions: 272 valid papers.
  • Overall Acceptance Rate: ~45% (123 papers).
  • Oral Presentations: 36 papers (13.2%).
  • Poster Presentations: 87 papers (32%).

In addition to the contributed papers, the conference featured keynote insights from three distinguished invited speakers: Katsushi Ikeuchi, Tomas Pajdla, and Yaser Ajmal Sheikh, who provided perspectives on the future of autonomous robotics and sensing.

Conclusion

3DV 2020 demonstrated the resilience of the 3D vision community in the face of global crisis. The successful management of a record-breaking submission pool and the transition to a virtual format underscore the field's vitality. As 3D vision continues to integrate with autonomous systems and advanced sensing, the rigorous standards established during this period provide a foundation for the conference's continued growth in the post-pandemic era.

Summary updated

BibTeX
@inproceedings{hilton2020message,
  author = {Adrian Hilton and Zuzana Kukelova and Stephen Lin and Jun Sato},
  title = {{Message from the 3DV 2020 Program Chairs}},
  booktitle = {3DV},
  year = {2020},
  doi = {10.1109/3dv50981.2020.00006},
}
Figure