Back to datasets
Dataset assetOpen Source CommunityImage Retrieval

SauravMaheshkar/roxford5k

This dataset is primarily used for image retrieval tasks and contains 10,126 training samples, each an image. Total size is 4,113,320,394.816 bytes; download size is 3,881,807,083 bytes. It is associated with a CVPR 2018 paper that revisits the Oxford and Paris large‑scale image retrieval benchmarks.

Source
hugging_face
Created
Nov 28, 2025
Updated
Jul 9, 2024
Signals
84 views
Availability
Linked source ready
Overview

Dataset description and usage context

Dataset Overview

Basic Information

  • License: Unknown
  • Dataset Size: 4,113,320,394.816 bytes
  • Download Size: 3,881,807,083 bytes
  • Papers with Code ID: image-retrieval-on-roxford-hard

Configuration

  • Config Name: default
  • Data Files:
    • Training: data/train-*

Features

  • Feature Name: image
  • Type: image

Splits

  • Training Set:
    • Samples: 10,126
    • Bytes: 4,113,320,394.816

Citation

@inproceedings{RITAC18,
  author = {Radenovi{c}, F. and Iscen, A. and Tolias, G. and Avrithis, Y. and Chum, O.},
  title = {Revisiting Oxford and Paris: Large-Scale Image Retrieval Benchmarking},
  booktitle = {CVPR},
  year = {2018}
}
Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio