extra_gated_prompt: > By requesting access, you agree to comply with the WONPIC-M50 Research Use License and the usage conditions described in this dataset card.

extra_gated_fields: Affiliation: text

Country: country

Intended use: type: select options:

Academic research
Education
Non-commercial personal project
Other

Project or research description: text

I agree to the WONPIC-M50 Research Use License: checkbox

I agree not to redistribute the original dataset: checkbox

I agree not to attempt to identify photographed individuals: checkbox

WONPIC-M50

Won’s Open-world Natural-scene Photographic Image Collection

WONPIC-M50 is a diverse travel photography dataset captured by Jiyun Won using a Canon EOS M50.

The dataset contains more than 1,000 photographs collected across multiple countries, cities, seasons, times of day, and visual environments. The images represent naturally occurring travel scenes rather than images collected under a controlled laboratory setup.

Overview

WONPIC-M50 was created to support research involving real-world photography and computational imaging.

Potential applications include:

Computational photography
Image restoration and enhancement
Image quality assessment
Camera and ISP analysis
Image geolocation
Visual recognition
Representation learning
Robustness and generalization evaluation

WONPIC-M50 is not designed as a task-specific benchmark and does not define an official training, validation, or test split in its initial release.

Dataset Characteristics

Creator: Jiyun Won
Camera: Canon EOS M50
Modality: Still photography
Image format: JPEG
Collection type: Real-world travel photography
Number of images: More than 1,000
Photographer count: One
Capture conditions: Uncontrolled real-world environments
Release version: 1.0
Release year: 2026

Image Content

The dataset includes a variety of travel-related scenes, which may include:

Natural landscapes
Urban environments
Streets and transportation
Buildings and architecture
Indoor and outdoor scenes
Daytime and nighttime photography
Images captured under different weather and lighting conditions
Everyday objects, food, signs, and cultural environments

The geographic and semantic distributions are not balanced.

Metadata

Depending on availability, the images may include metadata such as:

Capture date and time
Country, city, or region
Geographic coordinates
Camera model
Lens information
ISO sensitivity
Exposure time
Aperture
Focal length
Image dimensions

Some metadata fields may be missing for individual images. Users should not assume that every image contains complete or equally precise metadata.

Geographic information should be used responsibly and must not be used to infer sensitive personal locations or individual travel patterns.

Access

WONPIC-M50 is distributed as a gated dataset.

To request access:

Sign in to a Hugging Face account.
Complete the access request form.
Describe your affiliation and intended use.
Agree to the dataset license and usage conditions.
Wait for the request to be reviewed.

Access is granted to individual Hugging Face accounts and must not be shared with other individuals.

The dataset creator may approve, reject, or revoke access at their discretion.

Loading the Dataset

After receiving access, authenticate through the Hugging Face CLI:

hf auth login

The dataset can then be loaded with the Hugging Face datasets library:

from datasets import load_dataset

dataset = load_dataset(
    "jiyunwon/WONPIC-M50",
    split="train",
    token=True,
)

print(dataset)
print(dataset[0])

Users may also download the repository through the Hugging Face CLI:

hf download jiyunwon/WONPIC-M50 \
  --repo-type dataset \
  --local-dir WONPIC-M50

Permitted Uses

Subject to the terms in LICENSE.md, permitted uses include:

Non-commercial academic research
Educational use
Training and evaluating computer-vision models
Analysis of image formation, camera metadata, or photographic quality
Publication of aggregate experimental results
Publication of derived models that do not reproduce or redistribute the original images

Prohibited Uses

The following uses are prohibited:

Redistribution or public rehosting of the original images
Commercial use without separate written permission
Face recognition, identity inference, or person re-identification
Surveillance or profiling of photographed individuals
Inferring private or sensitive personal information
Harassment, discrimination, or other harmful applications
Claiming ownership of the original photographs
Using the dataset in violation of applicable laws or regulations

Privacy and Ethical Considerations

The images were captured in real-world travel environments and may incidentally include bystanders, vehicles, signs, artwork, buildings, or other third-party content.

Users must not attempt to:

Identify photographed individuals
Contact or track photographed individuals
Recover private information from visible content or metadata
Associate individuals with sensitive attributes
Use location metadata for surveillance or profiling

Users who discover sensitive, inappropriate, or personally identifying content should stop using the affected file and report it to the dataset creator.

See DATA_REMOVAL.md for removal and privacy-related requests.

Biases and Limitations

WONPIC-M50 reflects:

The travel history of a single photographer
The photographer’s preferences and framing choices
The characteristics of one camera system
Uneven geographic coverage
Uneven seasonal and temporal coverage
Uneven representation of scene categories
Camera-specific JPEG processing
Potential selection bias toward visually interesting scenes

The dataset should not be interpreted as a representative sample of the world, any country, population, culture, or demographic group.

Results obtained using this dataset may not generalize to other photographers, cameras, regions, or image-processing pipelines.

License

WONPIC-M50 is provided under the WONPIC-M50 Research Use License 1.0.

Review LICENSE.md before requesting access or using the dataset.

The photographs remain copyrighted by Jiyun Won. Access to the dataset does not transfer ownership of the images.

Citation

When using WONPIC-M50 in a publication, report, project, or derived dataset, cite it as follows:

@dataset{won2026wonpicm50,
  author    = {Won, Jiyun},
  title     = {WONPIC-M50: Won's Open-world Natural-scene Photographic Image Collection},
  year      = {2026},
  publisher = {Hugging Face},
  version   = {1.0}
}

Citation information may be updated when an associated paper or persistent DOI becomes available.

Version History

Version 1.0 — 2026

Initial public dataset card
Initial gated release
More than 1,000 Canon EOS M50 travel photographs
Associated photographic and location metadata where available

Contact

For dataset access, licensing, privacy, or removal inquiries, contact:

Jiyun Won Email: [add your contact email]

When reporting a specific image, include its file name and a brief description of the issue.

Downloads last month: 41

Total file size:

6.56 GB