0
点赞
收藏
分享

微信扫一扫

论文阅读 [TPAMI-2022] NCNet: Neighbourhood Consensus Networks for Estimating Image Correspondences

论文阅读 [TPAMI-2022] NCNet: Neighbourhood Consensus Networks for Estimating Image Correspondences

论文搜索(studyai.com)

搜索论文: NCNet: Neighbourhood Consensus Networks for Estimating Image Correspondences

搜索论文: http://www.studyai.com/search/whole-site/?q=NCNet:+Neighbourhood+Consensus+Networks+for+Estimating+Image+Correspondences

关键字(Keywords)

Feature extraction; Pattern matching; Task analysis; Convolutional neural networks; Electronic mail; Reliability; Benchmark testing; Neighbourhood consensus; geometric matching; image alignment; category-level matching

机器视觉

图像匹配; 立体匹配

摘要(Abstract)

We address the problem of finding reliable dense correspondences between a pair of images.

我们解决了在一对图像之间找到可靠的密集对应关系的问题。.

This is a challenging task due to strong appearance differences between the corresponding scene elements and ambiguities generated by repetitive patterns.

这是一项具有挑战性的任务,因为相应场景元素之间的外观差异很大,而且重复模式会产生歧义。.

The contributions of this work are threefold.

这项工作的贡献有三个方面。.

First, inspired by the classic idea of disambiguating feature matches using semi-local constraints, we develop an end-to-end trainable convolutional neural network architecture that identifies sets of spatially consistent matches by analyzing neighbourhood consensus patterns in the 4D space of all possible correspondences between a pair of images without the need for a global geometric model.

首先,受使用半局部约束消除特征匹配歧义的经典思想的启发,我们开发了一种端到端可训练卷积神经网络结构,该结构通过分析4D空间中一对图像之间所有可能对应的邻域一致性模式来识别空间一致性匹配集,而不需要全局几何模型。.

Second, we demonstrate that the model can be trained effectively from weak supervision in the form of matching and non-matching image pairs without the need for costly manual annotation of point to point correspondences.

其次,我们证明了该模型可以在弱监督下以匹配和非匹配图像对的形式进行有效训练,而无需昂贵的手动点对点对应标注。.

Third, we show the proposed neighbourhood consensus network can be applied to a range of matching tasks including both category- and instance-level matching, obtaining the state-of-the-art results on the PF, TSS, InLoc, and HPatches benchmarks…

第三,我们展示了所提出的邻里共识网络可以应用于一系列匹配任务,包括类别级和实例级匹配,在PF、TSS、InLoc和HPatches基准上获得最先进的结果。。.

作者(Authors)

[‘Ignacio Rocco’, ‘Mircea Cimpoi’, ‘Relja Arandjelović’, ‘Akihiko Torii’, ‘Tomas Pajdla’, ‘Josef Sivic’]

举报

相关推荐

0 条评论