CVPR Findings2026
EpiMask: Leveraging Epipolar Distance Based Masks in Cross-Attention for Satellite Image Matching
Rahul Deshmukh, Aditya Chauhan, Avinash Kak
Abstract
The deep-learning based image matching networks can now handle significantly larger variations in viewpoints and illuminations while providing matched pairs of pixels with sub-pixel precision. These networks have been trained with ground-based image datasets and, implicitly, their performance is optimized for the pinhole camera geometry. Consequently, you get suboptimal performance when such networks are used to match satellite images since those images are synthesized as a moving satellite camera records one line at a time of the points on the ground. In this paper, we present EpiMask, a semi-dense image matching network for satellite images that (1) Incorporates patch-wise affine approximations to the camera modeling geometry; (2) Uses an epipolar distance-based attention mask to restrict crossattention to geometrically plausible regions; and (3) That fine-tunes a foundational pretrained image encoder for robust feature extraction. Experiments on the SatDepth dataset demonstrate up to 30% improvement in matching accuracy compared to re-trained ground-based models. The code will be made available through epimask.git
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
- LightGlue: Local Feature Matching at Light Speed cites Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention
- LightGlue: Local Feature Matching at Light Speed cites DISK: Learning local features with policy gradient
Builds on11
- Swin Transformer: Hierarchical Vision Transformer using Shifted WindowsZe Liu, Yutong Lin, Yue Cao, Han Hu et al.ICCV 2021 · 31,683 citations
- LoRA: Low-Rank Adaptation of Large Language ModelsEdward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu et al.ICLR 2022 · 18,833 citations
- Transformers are RNNs: Fast Autoregressive Transformers with Linear AttentionAngelos Katharopoulos, Apoorv Vyas, Nikolaos Pappas, François FleuretICML 2020 · 2,665 citations
- LightGlue: Local Feature Matching at Light SpeedPhilipp Lindenberger, Paul-Edouard Sarlin, Marc PollefeysICCV 2023 · 936 citations
- DISK: Learning local features with policy gradientMichal J. Tyszkiewicz, Pascal Fua, Eduard TrullsNeurIPS 2020 · 652 citations
Related papers
- HomoMatcher: Achieving Dense Feature Matching with Semi-Dense Efficiency by Homography EstimationXiaolong Wang, Lei Yu, Yingying Zhang, Jiangwei Lao et al.AAAI 2025 · 6 citations
- Semi-Supervised Learning of Semantic Correspondence with Pseudo-LabelsJiwon Kim, Kwangrok Ryoo, Junyoung Seo, Gyuseong Lee et al.CVPR 2022 · 23 citations
- Multi-scale Matching Networks for Semantic CorrespondenceDongyang Zhao, Ziyang Song, Zhenghao Ji, Gangming Zhao et al.ICCV 2021 · 56 citations
- GLU-Net: Global-Local Universal Network for Dense Flow and CorrespondencesPrune Truong, Martin Danelljan, Radu TimofteCVPR 2020
- Sat2Density: Faithful Density Learning from Satellite-Ground Image PairsMing Qian, Jincheng Xiong, Gui-Song Xia, Nan XueICCV 2023 · 29 citations
