---
license: mit
task_categories:
- image-to-text
- image-segmentation
language:
- de
tags:
- table-structure-recognition
- receipt
size_categories:
- 1K<n<10K
pretty_name: kassenzettel-synth
---
# kassenzettel-synth

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

This dataset contains generate images of receipts.
It has been generated using [github.com/nimalu/kassenzettel](https://github.com/nimalu/kassenzettel).
The current version contains 1000 samples.

## Dataset Structure

For each instance, there is the underlying data, an image of the receipt, an image of the masks and both images augmented by adding crinkles.

![](receipt.png) 
![](masks.png)