Skip to content

Latest commit

 

History

History
29 lines (18 loc) · 726 Bytes

README.md

File metadata and controls

29 lines (18 loc) · 726 Bytes

DeepCD

You can download the pre-extracted patches and the pre-extracted features from https://drive.google.com/open?id=0B7c9IgGVCg6IVmUtY01Ga3hqWE0

You can download the pre-trained models from https://drive.google.com/open?id=0B7c9IgGVCg6INGp5OW5CcVhCTFU

usage

  1. dataProcess.m : extract patches from the Oxford dataset by VLFeat

    or you can directly use the pre-extracted patches

    dataProcess (run in matlab)
  1. extract.sh : extract features from patches by CNN-based models

    or you can directly use the pre-extracted features

    bash extract.sh
  1. result.m : evaluate the features extractes by CNN-based models
    result (run in matlab)