ubc-vision/image-matching-benchmark

how to evaluate on my own dataset?

Opened this issue · 1 comments

I would like to evaluate various features on my own dataset, I read the parts of the theory of generating scene data, but I cannot find the script to do this, and I don't know how those .h5 files are made.
Any clue of how to do this? Thanks.

Originally we didn't intend to run the scripts with other datasets. However, you can check this instructions:
https://github.com/ducha-aiki/creating-data-for-imc

If you don't use reality capture, skip that part