Official PyTorch code of DeepPanoContext: Panoramic 3D Scene Understanding with Holistic Scene Context Graph and Relation-based Optimization (ICCV 2021 Oral).
This work is so excellent! I'm curious about how to obtain the results of total3d and im3d, as the layout they generate is a cuboid. Also, whether I can train total3d and im3d with current codes, or may you share their results? I want to learn how to combine several single-view results into a panoramic one.
This work is so excellent! I'm curious about how to obtain the results of total3d and im3d, as the layout they generate is a cuboid. Also, whether I can train total3d and im3d with current codes, or may you share their results? I want to learn how to combine several single-view results into a panoramic one.