Problems and related results in algebraic vision and multiview geometry (Q6623823)

This article investigates the algebraic geometry underlying scene reconstruction in multiview vision, focusing on critical loci in static, dynamic, and segmented scenes. A key innovation is the use of Grassmann tensors to describe these loci, leveraging their properties such as rank, multi-rank, and kernel to better understand the associated varieties.\N\NThe authors make an important theoretical contribution by connecting Grassmann tensors with the ideals of critical loci, facilitating the identification of regions where reconstruction fails. This perspective is particularly valuable as it combines classical algebraic geometry with computational challenges in vision.\N\NMinor additions, such as examples and visualizations, could make the work more accessible and impactful.