Caption: 

This scientific poster titled "GenNBV: Generalizable Next-Best-View Policy for Active 3D Reconstruction" by Xiao Chao, Quanyi Li, Tianfu Xu, Ti Wang, Jiangmiao Pang from OpenGVDBLab, Shanghai AI Laboratory, and Hong Kong University of Science and Technology, details innovative research on 3D reconstruction. Key sections include:

1. **Motivation:** Highlights challenges with existing NBV policies, which include limited generalization ability and requiring manual parameter tuning, and proposes GenNBV as a solution.
2. **Objectives and Approach:** Describes the objectives of designing a versatile NBV policy that requires minimal human intervention and can generalize well across different 3D scenes.
3. **Methodology:** Explains the policy pipeline, which includes training phase, policy generation, and usage phase, aimed at improving efficiency and accuracy in 3D reconstruction processes.
4. **Visualization Comparisons:** Compares the reconstruction quality of different NBV policies, showing the superiority of GenNBV with various visual and quantitative metrics.
5. **Generalization Evaluation:** Demonstrates GenNBV’s ability to handle diverse 3D environments and objects, illustrated by coverage tables and graphs.
6. **Projects and Contact Information:** Provides QR codes for further project details and contact information for lead researchers.

The poster is a comprehensive exhibit of the team's advancements in creating a scalable and robust next-best-view policy for active 3D reconstruction, aimed at enhancing the field of computer vision and robotics.
Text transcribed from the image:
CATLAW
Motivation
Exing Nity
GenNBV: Generalizable Next-Best-View Policy for Active 3D Reconstruction
Xiao Chan Quany L Tal Wang Tentan Xue Jangnias Pang
Opetab Shangha Labory
The Orem Usersity of Hong Kong
Methodology
+ Osrview
+ Project Page AR
+ Contact GRANAMAN
Yer Capturing
Kay s
ability Evaluates
Ablation Stody