Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Matchable Image Retrieval by Learning from Surface Reconstruction #269

Open
chullhwan-song opened this issue Dec 26, 2019 · 1 comment
Open

Comments

@chullhwan-song
Copy link
Owner

chullhwan-song commented Dec 26, 2019

https://arxiv.org/abs/1811.10343
https://github.com/hlzz/mirror

@chullhwan-song
Copy link
Owner Author

Abstract

  • image retrieval > 먼저 새로운 개념이 나와서 이것만. 리뷰
  • 3D reconstruction

Method

Pre-Matching Regional Code (PRC)

  • 기본적으로 Deep Feature에 대한 새로운 개념은 없다.
  • R-MAC, SPoc, L2 Pooling(Max결과에 sqrt > 소스 참조)개념중에 특히, R-MAC를 Base한다.
  • 여기서는 기본적으로 Region별 vector를 합치는게 문제가 될수 있다고 보고 있음. > the mixed regional information may weaken its expressive power.
  • 이 문제의 해결책으로 "pre-matching regional code (PRC)" 개념을 제안. > R-MAC처럼 적용한다고 해서 "PR-MAC"
  • 사실 큰 개념은 아니고 Region 별 vector를 독립적으로 보고 matching한다는 개념 -> 즉, 한 이미지에 여러개의 vector가 존재하게 되는거고, 당연히 matching cost를 늘어갈것임.
    image
    • query image image의 region 중의 하나를 vector image
    • target 이미지 image의 region별 vector set image
  • 최종적인 distnace는,
    image
  • 하지만 언급했듯이 matching cost가 많이드니, 먼저 R-MAC를 이용하고, Re-Ranking용으로 PRE-MAC를 적용한다.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant