Visual alignment, also known as the correspondence or registration problem.
optical flow, 3D matching, medical imageing, tracking and augmented reality.
Many works has been made on pairwise alignment (aligning images A to image B)
-
FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks
Eddy Ilg, Nikolaus Mayer, Tonmoy Saikia, Margret Keuper, Alexey Dosovitskiy, Thomas Brox
[CVPR 2017] -
Convolutional neural network architecture for geometric matching
Ignacio Rocco, Relja Arandjelović, Josef Sivic
[CVPR 2017] -
End-to-end weakly-supervised semantic alignment
Ignacio Rocco, Relja Arandjelović, Josef Sivic
[CVPR 2018] -
Neighbourhood Consensus Networks
Ignacio Rocco, Mircea Cimpoi, Relja Arandjelović, Akihiko Torii, Tomas Pajdla, Josef Sivic
[NeurIPS 2018] -
Attentive Semantic Alignment with Offset-Aware Correlation Kernels
Paul Hongsuck Seo, Jongmin Lee, Deunsol Jung, Bohyung Han, Minsu Cho
[ECCV 2018] -
Hyperpixel Flow: Semantic Correspondence with Multi-layer Neural Features
Juhong Min, Jongmin Lee, Jean Ponce, Minsu Cho
[ICCV 2019] -
RAFT: Recurrent All-Pairs Field Transforms for Optical Flow
Zachary Teed, Jia Deng
[ECCV 2020] -
Learning to Compose Hypercolumns for Visual Correspondence
Juhong Min, Jongmin Lee, Jean Ponce, Minsu Cho
[ECCV 2020] -
Semantic Correspondence as an Optimal Transport Problem
Yanbin Liu, Linchao Zhu, Makoto Yamada, Yi Yang
[CVPR 2020] -
Learning to Compose Hypercolumns for Visual Correspondence
Juhong Min, Jongmin Lee, Jean Ponce, Minsu Cho
[ECCV 2020] -
CATs: Cost Aggregation Transformers for Visual Correspondence
Seokju Cho, Sunghwan Hong, Sangryul Jeon, Yunsung Lee, Kwanghoon Sohn, Seungryong Kim
[NeurIPS 2021] -
Convolutional Hough Matching Networks
Juhong Min, Minsu Cho
[CVPR 2021]
The problem of global joint alignment has not received as much attention.
Training on jointly aligned datasets can produce higher quality generative models than training on unaligned data.
Based on Congealing method
Data driven image models through continuous joint alignment.