CornerFormer: 提高角点表示以进行细粒度结构重建【英文版】

英文标题：CornerFormer: Boosting Corner Representation for Fine-Grained Structured Reconstruction中文摘要：本论文提出了一种改进的角落表征方法，将角点检测和边缘预测之间的知识融合，建议角点候选者在四个 heatmap 通道中提出，定量和定性评估证明该方法能够更好地重建细粒度结构，并取得了比最先进的 Corner 和 Edge 模型更好的表现。英文摘要：Structured reconstruction is a non-trivial dense prediction problem, whichextracts structural information (\eg, building corners and edges) from a rasterimage, then reconstructs it to a 2D planar graph accordingly. Compared withcommon segmentation or detection problems, it signi