🎉 Exciting News! My Paper Accepted at ICCV 2023’s R6D Workshop! 📚📊

I am thrilled to share that my paper titled “Reconstruction of 3D Interaction Models from Images using Shape Prior” has been accepted at the 8th International Workshop on Recovering 6D Object Pose (R6D) as part of ICCV 2023! 🌟🔍

🔬 Abstract: We investigate the reconstruction of 3D human-object interactions from images, encompassing 3D human shape and pose estimation as well as object shape and pose estimation. To address this task, we introduce an autoregressive transformer-based variational autoencoder capable of learning a robust shape prior from extensive 3D shape datasets. Additionally, we leverage the reconstructed 3D human body as supplementary features for object shape and pose estimation. In contrast, prior methods only predict object pose and rely on shape templates for shape prediction. Experimental findings on the BEHAVE dataset underscore the effectiveness of our proposed approach, achieving a 40.7cm Chamfer distance and demonstrating the advantages of learning a shape prior.

🙏 Acknowledgments: I extend my heartfelt gratitude to Yen-Ling Kuo (Assistant Professor at the University of Virginia) and Xi Wang (Postdoc Fellow in AIT lab ETH and Postdoc researcher in CV lab ETH) for their invaluable support and collaboration.

🌍 ICCV 2023: While I won’t be able to attend the conference in person, I’m genuinely excited to contribute to the workshop’s discussions and connect with fellow researchers in the field.

This marks my first accepted paper, and I couldn’t be more thrilled about this achievement! 🚀 Thank you to everyone who has been a part of this journey.