2D Gum-Net：无监督稠密弹性图像配准新方法
2D Gum-Net: Unsupervised Dense Elastic Image Registration via Symmetric Correlation and Spectral Pooling
做医学影像、指纹匹配或任何非刚性配准的人都该看看这篇。它把 Gum-Net 从 3D 冷冻电镜搬到 2D，全程不用标注形变场，谱池化还带来约 50 倍 GPU 加速，在 FVC2004 上零样本就有 15.7% 的提升。
arXiv 论文提出 2D Gum-Net，将原用于冷冻电子断层扫描 3D 子断层对齐的 Geometric Unsupervised Matching Network 迁移到 2D 弹性配准，实现端到端全无监督的逐像素形变场预测，无需真实形变场标注。技术上，三个模块全部重设计为 2D 版本，刚体变换头换成 Dense Spatial Transformer，回归稀疏控制点网格再经双三次插值得到全分辨率形变场 Φ；下采样采用 DCT 谱池化，以冻结基矩阵乘法实现，带来约 50 倍 GPU 加速，保留最大/平均池化会破坏的细粒度准周期结构。
评估选在接触式指纹这一高难度场景：局部自相似的脊线让对应关系高度歧义，皮肤按压传感器产生真实的大位移弹性形变，细节点则作为精确地标供独立匹配器外部验证形变场的拓扑正确性。模型仅用合成指纹训练，在真实基准上零样本评估，FVC2004 上结构相关性提升最高 15.7%，独立匹配器下的真匹配分数上升而假匹配分数基本不变，说明预测形变恢复了真实结构而非伪造对应关系。
2D Gum-Net: Unsupervised Dense Elastic Image Registration via Symmetric Correlation and Spectral Pooling
Dense non-rigid registration, the recovery of a pixel-wise deformation field that warps one image into geometric agreement with another, is fundamental across medical imaging, structural biology, and correspondence estimation, yet learning-based approaches face a severe bottleneck: ground-truth deformation fields are notoriously difficult to obtain for real data. We present 2D Gum-Net, an end-to-end architecture that predicts dense elastic deformation fields fully unsupervised. We transfer the Geometric Unsupervised Matching Network (Gum-Net), originally developed for 3D subtomogram alignment in cryo-electron tomography, to 2D elastic registration: all three modules are redesigned in 2D, and the rigid transformation head is replaced by a Dense Spatial Transformer that regresses a sparse control-point grid, lifted to a full-resolution field $Φ$ by bicubic interpolation. Downsampling uses DCT spectral pooling, implemented as frozen-basis matrix multiplications ($\approx 50\times$ GPU speedup), preserving the fine, quasi-periodic structure that max and average pooling destroy. We evaluate on contact-based fingerprints, a challenging testbed for this task: locally self-similar ridge patterns make correspondence maximally ambiguous, skin pressed against a sensor generates authentic large-displacement elastic distortion, and minutiae act as precise landmarks through which an independent matcher externally verifies the topological correctness of predicted fields. Trained purely on synthetic impressions and evaluated zero-shot on real benchmarks, our method improves structural correlation by up to 15.7% on FVC2004 and raises genuine match scores under an independent matcher while leaving imposter scores essentially unchanged. This suggests that the predicted deformations restore true structure without fabricating correspondences.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Gum-Net, 图像配准, 计算机视觉, 无监督学习.
Source attribution: curated from arXiv cs.AI (@Aarya Bhave, Anushka Ajay Bhave).
