HyperAI

Oxford105k Oxford Architectural Image Dataset

Date

3 years ago

Size

7.17 GB

Organization

University of Oxford

License

其他

Categories

特色图像

The Oxford105k dataset consists of 5,062 images that were obtained by searching for specific Oxford landmarks on Flicker. The dataset is manually annotated to generate comprehensive ground truth for 11 different landmarks, each with 5 relevant query results, for a total of 55 query results. This dataset is used to evaluate search performance for large-scale object retrieval (such as mAP).

Oxford105k.torrent
Seeding 2Downloading 1Completed 73Total Downloads 129
  • Oxford105k/
    • README.md
      1.12 KB
    • README.txt
      2.24 KB
    • README2.txt
      105.89 KB
    • compute_ap.cpp
      107.63 KB
      • data/
        • README2.txt
          211.27 KB
        • face-annotations-v1.json
          1.01 MB
        • feat_oxc1_hesaff_sift.bin.tgz
          1.5 GB
        • gt_files_170407.tgz
          1.5 GB
        • oxbuild_images-v1.tgz
          3.3 GB
        • word_oxc1_hesaff_sift_16M_1M.tgz
          3.59 GB
    • face-annotations-v1.json
      3.59 GB
    • feat_oxc1_hesaff_sift.bin.tgz
      5.09 GB
    • gt_files_170407.tgz
      5.09 GB
    • oxbuild_images-v1.tgz
      6.89 GB
    • word_oxc1_hesaff_sift_16M_1M.tgz
      7.17 GB