Researchers have introduced Raster2Seq, a novel method for reconstructing structured vector-graphics representations from rasterized floorplan images. This approach frames floorplan reconstruction as a sequence-to-sequence task, where rooms, windows, and doors are represented as labeled polygon sequences that encode both geometry and semantics. An autoregressive decoder predicts the next corner based on image features and previously generated corners, guided by learnable anchors. Raster2Seq demonstrates state-of-the-art performance on benchmarks like Structure3D and CubiCasa5K, and shows strong generalization to more complex datasets such as WAFFLE. AI
IMPACT This method could improve automated understanding and CAD workflows for complex indoor spaces.
RANK_REASON The cluster contains a research paper detailing a new method for floorplan reconstruction. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
