Floorplan Reconstruction

Floorplan reconstruction aims to automatically generate accurate 2D floorplan representations from various input data, such as 3D point clouds, images, or sparse panoramas, primarily focusing on improving accuracy and efficiency. Current research heavily utilizes deep learning models, including transformers and graph neural networks, often incorporating room-wise representations and geometric constraints to enhance the quality and plausibility of reconstructed floorplans. This field is significant for its potential to automate tasks in architecture, construction, and robotics, streamlining workflows and improving efficiency in building information modeling and spatial understanding.