Is there a plan to release rest of evaluations?

[msharmavikram]

Appreciate in making this dataset available.

I was wondering if you plan to release the other evaluations codebase to the community. Currently, I see that R-[37]-GeM is released. Is there any plan to release other?

If not, perhaps, can you do release DELF-ASMK+SP, HesAff-rSIFT-ASMK+SP, R-[10]-R-MAC+DFS?

[filipradenovic]

Most of the methods benchmarked in this project are available online:

• R-[37]-GeM: https://github.com/filipradenovic/cnnimageretrieval & https://github.com/filipradenovic/cnnimageretrieval-pytorch
• R-[10]-R-MAC: http://www.europe.naverlabs.com/Research/Computer-Vision/Learning-Visual-Representations/Deep-Image-Retrieval
• DELF: https://github.com/tensorflow/models/tree/master/research/delf
• ASMK*: http://cmp.felk.cvut.cz/~toliageo/soft.html
• HesAff: https://github.com/perdoch/hesaff & https://github.com/ducha-aiki/hesaff-pytorch (both contain SIFT implementation as well I believe)
• rSIFT: https://github.com/ducha-aiki/numpy-sift & https://github.com/ducha-aiki/pytorch-sift
• DFS: https://github.com/ahmetius/diffusion-retrieval & https://github.com/ducha-aiki/manifold-diffusion

I might have skipped or do not know of other implementations of the same methods. I am sure there are many more.

For now, we do not plan to release the unified evaluation of all the methods. Features for R-[37]-GeM are added to the code for the example evaluation purpose.

[msharmavikram]

Thanks! this helps to get started.

[Varshul]

Hey @filipradenovic. Great work on the paper and sharing of individual features.

Though for further evaluations would really be amazing if you can release features of DELF-ASMK*+SP (primarily) and others.

[filipradenovic]

This is a new implementation of DELF-ASMK* available: github.com/gtolias/asmk. Additionally, there is a new implementation of HQE available: github.com/gtolias/hqe

For now, we are not planning to release all the methods used in the paper, as many of them are already available online.