通过球面归一化提升图像关键点匹配精度与效率
Normalized Matching Transformer
- 采用球面归一化与对比损失，增强特征区分度
- 在PascalVOC和SPair-71k上超越现有方法5.1%以上
- 训练收敛速度比主流模型快1.7倍，适合实时应用
我们提出一种名为归一化匹配变压器（Normalized Matching Transformer, NMT）的深度学习方法，用于高效准确地进行图像对之间的稀疏语义关键点匹配。NMT 包含一个强大的视觉主干网络，通过 SplineCNN 进行几何特征精炼，随后使用归一化 Transformer 计算匹配特征。其核心是超球面归一化策略：在每一层 Transformer 中强制嵌入向量为单位范数，并结合对比 InfoNCE 损失与超球面均匀性损失进行训练，从而获得更具判别力的关键点表示。该架构/损失组合促使匹配特征在每一层都实现紧密对齐，非匹配特征则保持较大距离。尽管结构简单，NMT 在 PascalVOC 与 SPair-71k 数据集上均达到新基准，相比 BBGM、ASAR、COMMON 与 GMTR 分别提升 5.1% 和 2.2%，且收敛速度至少快 1.7 倍。
原文摘要 · Abstract (English)
We introduce the Normalized Matching Transformer (NMT), a deep learning approach for efficient and accurate sparse semantic keypoint matching between image pairs. NMT consists of a strong visual backbone, geometric feature refinement via SplineCNN, followed by a normalized Transformer for computing matching features. Central to NMT is our hyperspherical normalization strategy: we enforce unit-norm embeddings at every Transformer layer and train with a combined contrastive InfoNCE and hyperspherical uniformity loss to yield more discriminative keypoint representations. This novel architecture/loss combination encourages close alignment of matching image features and large distances between non-matching ones not only at the output level, but for each layer. Despite its architectural simplicity, NMT sets a new state-of-the-art performance on PascalVOC and SPair-71k, outperforming BBGM, ASAR, COMMON and GMTR by 5.1% and 2.2%, respectively, while converging in at least 1.7x fewer epochs compared to other state-of-the-art baselines. These results underscore the power of combining pervasive normalization with hyperspherical learning for matching tasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。