README.md

---
dataset_info:
  features:
  - name: Flickr ID
    dtype: int64
  - name: url
    dtype: string
  - name: owner
    dtype: string
  - name: Make
    dtype: string
  - name: Model
    dtype: string
  - name: Exposure
    dtype: string
  - name: Aperture
    dtype: string
  - name: ISO Speed
    dtype: string
  - name: Focal Length
    dtype: string
  splits:
  - name: train
    num_bytes: 51182479
    num_examples: 356459
  download_size: 13466120
  dataset_size: 51182479
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: cc-by-4.0
---

# Dataset Card for Dataset Name

<!-- Provide a quick summary of the dataset. -->

FlickrExif provides images and their corresponding acquisition labels (e.g. camera model, aperture) extracted from Exif metadata.

## Dataset Details

### Dataset Description

<!-- Provide a longer summary of what this dataset is. -->



- **Curated by:** Ryan Ramos, Vladan Stojnic, Giorgos Kordopatis-Zilos, Yuta Nakashima, Giorgos Tolias, Noa Garcia
- **Funded by:** Dataset collection was supported by JSPS KAKENHI No. JP23H00497 and JP22K12091, JST CREST Grant No. JPMJCR20D3, and JST FOREST Grant No. JPMJFR216O.
<!-- - **Shared by [optional]:** [More Information Needed] -->
<!-- - **Language(s) (NLP):** [More Information Needed] -->
- **License:**  [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)

### Dataset Sources

<!-- Provide the basic links for the dataset. -->

- **Repository:** [https://github.com/ryan-caesar-ramos/visual-encoder-traces](https://github.com/ryan-caesar-ramos/visual-encoder-traces)
- **Paper:** [Processing and acquisition traces in visual encoders: What does CLIP know about your camera?](https://arxiv.org/abs/2508.10637)
<!-- - **Demo [optional]:** [More Information Needed] -->
- **Summary thread**: [by Vladan Stojnić, co-first author](https://bsky.app/profile/stojnicv.xyz/post/3lwo7xswiu22n)

## Uses

<!-- Address questions around how the dataset is intended to be used. -->

### Direct Use

<!-- This section describes suitable use cases for the dataset. -->

This dataset is intended to be used for probing visual encoders' abilities to encode acquisition labels into their embeddings.

### Out-of-Scope Use

<!-- This section addresses misuse, malicious use, and uses that the dataset will not work well for. -->

Due to potential correlations between image semantics and acquisition labels, we do not recommend using this dataset to evaluate acquisition label prediction without preprocessing (e.g. masking, as done in the paper this dataset accompanies). Additional, due to limitations described below, we do not endorse the use of this dataset for tasks where balanced data is extremely important (e.g. image generation).

## Dataset Structure

<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->

- Flickr ID: the identifier of the image on Flickr
- url: URL to the Flickr image
- owner: identifier of the account that uploaded the image to Flickr
- Make: the manufacturer of the camera used to capture the image
- Model: the camera model used to capture the image
- Exposure: the amount of time that light that was allowed to enter the camera while capturing the image
- Aperture: the size of the opening in the lens and the corresponding amount of light thus allowed to enter the camera while capturing the image
- ISO Speed: the sensitivity to light of the camera sensor used to capture the image
- Focal Length: the distance between the center of the camera's lens and the camera's sensor

If your goal is to use this dataset to recreate our results, note that there is no single train set. We create splits per attribute, derive new attributes, and manually intervene on some labels. To recreate our results, please use the JSON files under `split_data/`. You can use the following snippet to download them:

```
from huggingface_hub import snapshot_download

snapshot_download(
  repo_id="ryanramos/FlickrExif",
  allow_patterns="split_data/*.json",
  repo_type="dataset"
)
```

## Dataset Creation

### Curation Rationale

<!-- Motivation for the creation of this dataset. -->

Many datasets to do not include Exif metadata, and those that include them to do not cover acquisition parameters such as camera model. We close this gap with FlickrExif.

### Source Data

<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->

#### Data Collection and Processing

<!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->

We used the FlickrAPI to search for 2,000-4,000 safe-for-work, permissively licensed images each month from January 2000 to August 2025. To prevent the dataset from being dominated by prolific photographers, we limit the number of images per photographer to 10 per month and year.

#### Who are the source data producers?

<!-- This section describes the people or systems who originally created the data. It should also include self-reported demographic or identity information for the source data creators if this information is available. -->

This data is collected from Flickr users who marked their images are safe-for-work and permissively licensed, according to Flickr API's search filters.

<!-- ### Annotations [optional] -->

<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->

<!-- #### Annotation process -->

<!-- This section describes the annotation process such as annotation tools used in the process, the amount of data annotated, annotation guidelines provided to the annotators, interannotator statistics, annotation validation, etc. -->

<!-- #### Who are the annotators? -->

<!-- This section describes the people or systems who created the annotations. -->

#### Personal and Sensitive Information

<!-- State whether the dataset contains data that might be considered personal, sensitive, or private (e.g., data that reveals addresses, uniquely identifiable names or aliases, racial or ethnic origins, sexual orientations, religious beliefs, political opinions, financial or health data, etc.). If efforts were made to anonymize the data, describe the anonymization process. -->

Flickr user IDs are included in this dataset which can be used to access their profile pages.

## Bias, Risks, and Limitations

<!-- This section is meant to convey both technical and sociotechnical limitations. -->

There may exist correlations between an image's semantic content and its acquisition labels. For example, nighttime photos may be associated with higher ISO values. This can potentially affect acquisition label prediction if models rely on semantic cues as shortcuts. Furthermore, as this dataset was created with a specific use case in mind, the only filtering or balancing used to create this dataset is the owner filtering process described in the in the Data Collection and Processing section. Any remaining harmful content or imbalance resulting from the API's search algorithm are left untouched.

### Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

In acquisition label prediction settings, we highly recommend users to scrub semantic content from images with techniques such as masking, as done in the paper this dataset accompanies. Furthermore, because of the additional limitations mentioned previously, we do not endorse the use of FlickrExif for use cases outside its intended one.

## Citation

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**

```
@InProceedings{Ramos_2025_ICCV,
    author    = {Ramos, Ryan and Stojni\'c, Vladan and Kordopatis-Zilos, Giorgos and Nakashima, Yuta and Tolias, Giorgos and Garcia, Noa},
    title     = {Processing and acquisition traces in visual encoders: What does CLIP know about your camera?},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2025},
    pages     = {17056-17066}
}
```

**APA:**

Ramos, R., Stojnić, V., Kordopatis-Zilos, G., Nakashima, Y., Tolias, G., & Garcia, N. (2025). Processing and acquisition traces in visual encoders: What does CLIP know about your camera?. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) (pp. 17056-17066).

<!-- ## Glossary [optional]　-->

<!-- If relevant, include terms and calculations in this section that can help readers understand the dataset or dataset card. -->

<!-- [More Information Needed]

## More Information [optional]

[More Information Needed] -->

## Dataset Card Authors

Ryan Ramos

## Dataset Card Contact

ryanramos@is.ids.osaka-u.ac.jp