---
license: apache-2.0
pretty_name: Image Transformation Sequences
tags:
  - video
  - computer-vision
  - multimodal
---

# Image Transformation Sequences

### 🎬 CLICK THE PREVIEW IMAGE BELOW TO OPEN THE VIDEO

[![Transformation Sequence Preview](PBv1trains.png)](videos/three_panel_repeated_shuffle.webm)

Click the preview image above to open the example WebM video.

A collection of video sequences showing photographs evolving through structured visual transformations over time.

Each example contains:

* Original source imagery
* A transformation sequence rendered as video
* Multiple stages of visual evolution within a single clip
* Self-contained examples suitable for viewing, analysis, and research

## Dataset Overview

This dataset is designed for studying visual change over time from a fixed source image set.

Each video begins with original imagery and progresses through a sequence of evolving visual states before reaching a transformed final composition.

The dataset focuses on:

* Temporal visual evolution
* Image-to-sequence relationships
* Multi-stage transformation processes
* Video-based representation learning
* Visual change analysis

## Format

Each record contains:

* Video (WebM)
* Optional caption metadata
* Embedded source imagery within the video presentation

## Potential Uses

* Video understanding
* Temporal representation learning
* Transformation tracking
* Sequence modeling
* Visual retrieval
* Dataset visualization
* Human evaluation studies

## Dataset Structure

```text
videos/
├── sample_0001.webm
├── sample_0002.webm
├── sample_0003.webm
└── ...

metadata.jsonl
README.md
LICENSE
```

Example JSONL record:

```json
{
  "video": "videos/sample_0001.webm",
  "caption": "Image transformation sequence"
}
```

## Notes

The dataset is intended as a collection of transformation sequences rather than a collection of independent images.

Source imagery and transformed states are presented together within each video to provide visual context while maintaining a compact, self-contained format.

The repository focuses on the resulting transformation sequences and does not require access to the production workflow used to create them.

## License

This dataset is released under the Apache License 2.0.

You are free to:

* Use the dataset for research and commercial purposes
* Modify and adapt the dataset
* Distribute copies and derivative works
* Incorporate the dataset into larger projects and workflows

Users must comply with the terms of the Apache 2.0 license, including preservation of required notices and attribution where applicable.

For full license details, see the LICENSE file included with this repository.

## Citation

If you use this dataset in research, publications, benchmarks, or derivative works, please cite the repository and the dataset version used.