IEEE RA-L

Diff2DGS: Reliable Reconstruction of Occluded Surgical Scenes via 2D Gaussian Splatting

Diff2DGS: 通过2D高斯点散布实现遮挡手术场景的可靠重建

Tianyi Song, Danail Stoyanov, Evangelos Mazomenos, Francisco Vasconcelos

AI总结 Diff2DGS通过两阶段框架提升遮挡手术场景的3D重建精度，结合扩散模型和可学习变形模型，实现更准确的几何和外观重建。

Comments This work has been accpeted by the IEEE Robotics and Automation Letters(RA-L)

MSG-Loc: Multi-Label Likelihood-based Semantic Graph Matching for Object-Level Global Localization

MSG-Loc: 基于多标签似然的语义图匹配用于物体级全局定位

Gihyeon Lee, Jungwoo Lee, Juwon Kim, Young-Sik Shin, Younggun Cho

机构 * Electrical and Computer Engineering, Inha University（信息与通信技术规划与评估机构，延世大学） ； Depart. of AI Machinery, Korea Institute of Machinery and Materials（人工智能机械系，韩国机械材料研究院） 纠错

AI总结 本文提出基于多标签似然的语义图匹配框架，用于解决物体级全局定位中的语义模糊性和误分类问题。

Comments Accepted in IEEE Robotics and Automation Letters (2025)