zzcvdnn is a C++ library that extend OpenCV DNN module to do keypoint detection, descriptor extraction and matching.
- SuperPoint, CVPRW 2018
- DISK, NeurIPS 2020
- SGMNet, ICCV 2021
- LANet, ACCV 2022
- ALIKE, TMM 2022
- ALIKED, ArXiv 2023: Use deformable convolution
- SFD2, CVPR 2023: Use SuperPoint class & NN Matcher
- silk, ArXiv 2023: Fused BN and Conv
- DeDoDe 🎶, 3DV 2024
- Steerers, arXiv 2023: Based on DeDoDe
- XFeat, CVPR 2024: Fused Linear and BN
- Patch2Pix, CVPR 2021
- LoFTR, CVPR 2021: Use gim weights
- SE2-LoFTR, CVPRW 2022: Use e2cnn
- ASpanFormer, ECCV 2022
- MatchFormer, ACCV 2022
- TopicFM, AAAI 2023
- TopicFM+, arXiv 2023
- AdaMatcher, CVPR 2023
- Efficient LoFTR, CVPR 2024
- SAM-Net, ESWA 2023
- DenseMatching
- RoRD, IROS 2021
- DKM, CVPR 2023: Use
torch.linalg.inv - Roma&TinyRoma, CVPR 2024: Use
torch.linalg.inv
- NearestNeighborMatcher
- DualSoftMaxMatcher
- SuperGlue, CVPR 2020
- LightGlue, ICCV 2023
- GlueStick, ICCV 2023
- OmniGlue, CVPR 2024
- glue-factory
- gim, ICLR 2024
- https://github.com/Vincentqyw/image-matching-webui
- https://github.com/Vincentqyw/LineSegmentsDetection
- https://github.com/gmberton/image-matching-models
- https://github.com/3DOM-FBK/deep-image-matching
- https://github.com/ericzzj1989/Awesome-Image-Matching
- https://github.com/chicleee/Image-Matching-Paper-List
%%{init: {'theme':'forest'}}%%
mindmap
root((zzcvdnn))
FeatureExtractor
(ALike)
(ALiked)
(DeDoDe)
(DISK)
(LANet)
(SiLK)
(SuperPoint)
(SuperPointDino)
(SPWireframe)
(SFD2)
(XFeat)
GlueMatcher
(SuperGlue)
(LightGlue)
(OmniGlue)
GlueStick
FusedMatcher
(OnnxFusedMatcher)
{{LoFTR}}
{{SE2-LoFTR}}
{{ELoFTR}}
{{ASpanFormer}}
{{MatchFormer}}
{{SAM-Net}}
(RoMatcher)
{{DKM}}
{{RoMa}}
{{Tiny-RoMa}}
DescriptorsMatcher
(NearestNeighborMatcher)
(DualSoftMaxMatcher)
(RotationSteerersMatcher)
Images
