- 2024-08-09Local All-Pair Correspondence for Point Tracking 中英对照
论文来自：https://ku-cvlab.github.io/locotrack/LocalAll-PairCorrespondenceforPointTracking局部全对应对点跟踪SeokjuCho\({}^{1}\),JiahuiHuang\({}^{2}\),JisuNam\({}^{1}\),HonggyuAn\({}^{1}\),Seungryong\({\mathrm{{Kim}}}^{1,\dagger}\),
- 2023-11-13Cost Aggregation with Transformers for Sparse Correspondence-读书笔记
CostAggregationwithTransformersforSparseCorrespondence:2022背景:该论文结合了SuperGlue和CATs，将里面所有手工制作的部分都代替了。将CATs引入该模型，用Transformer取代手工制作的成本聚合方法，用于具有自关注层全局接受域的可学习成本聚合。（PS：成本聚合：成本聚合是指在立
- 2023-11-12A Learning Method for Feature Correspondence with Outliers读书笔记
ALearningMethodforFeatureCorrespondence withOutliers------2022年论文地址：ALearningMethodforFeatureCorrespondencewithOutliers|IEEEConferencePublication|IEEEXplore背景：SuperGlue在许多现实任务中取得了优异的性能，但在处理离群特征方面仍然存在问
- 2023-10-04Q-REG论文阅读
Q-REGJin,S.,Barath,D.,Pollefeys,M.,&Armeni,I.(2023).Q-REG:End-to-EndTrainablePointCloudRegistrationwithSurfaceCurvature.paper:2309.16023v1]Q-REG:End-to-EndTrainablePointCloudRegistrationwithSurfaceCurvature(arxiv.org)