infinite-dataset-hub/InpaintingData

---
license: mit
tags:
- infinite-dataset-hub
- synthetic
---

# InpaintingData

tags: image restoration, supervised learning, machine vision

_Note: This is an AI-generated dataset so its content may be inaccurate or false_

**Dataset Description:** The 'InpaintingData' dataset is curated for researchers and practitioners in the field of computer vision, specifically focusing on the task of image inpainting using supervised learning techniques. Image inpainting refers to the process of reconstructing lost or deteriorated parts of images and videos. This dataset comprises various images with artificially introduced occlusions (holes) and their corresponding ground truth images where the occlusions have been filled in. The labels in the dataset are based on the types of objects and contexts that are being restored, such as 'natural', 'urban', 'animals', 'people', and 'artifacts'.

**CSV Content Preview:**

```

id,filename,occurrence,label,occlusion_type,restoration_success

001,image001.jpg,1,natural,tree,high

002,image002.jpg,1,urban,building,medium

003,image003.jpg,1,animals,dog,high

004,image004.jpg,1,people,person,medium

005,image005.jpg,1,artifacts,artifact,low

```

**Source of the data:**

The dataset was generated using the [Infinite Dataset Hub](https://huggingface.co/spaces/infinite-dataset-hub/infinite-dataset-hub) and microsoft/Phi-3-mini-4k-instruct using the query 'inapaint':

- **Dataset Generation Page**: https://huggingface.co/spaces/infinite-dataset-hub/infinite-dataset-hub?q=inapaint&dataset=InpaintingData&tags=image+restoration,+supervised+learning,+machine+vision
- **Model**: https://huggingface.co/microsoft/Phi-3-mini-4k-instruct
- **More Datasets**: https://huggingface.co/datasets?other=infinite-dataset-hub