---
license: unknown
task_categories:
- image-classification
size_categories:
- 1K<n<10K
---

# MV-VDB-photos-small

**Media Vault - Vector Database Photos (Small)**

A curated collection of 11,000 images from various computer vision datasets, designed for testing internal mechanisms in the Media Vault Vector Database system. This is the first small-scale dataset (targeting 10K samples, with NSFW split totaling 11K) for validation and testing purposes.

## Dataset Structure

The dataset contains two splits:
- **`sfw`**: All non-NSFW images (~10,000 images)
- **`x_nsfw`**: Only NSFW images (~1,000 images)

Each image has the following columns:
- `index`: Image index within the dataset
- `image`: Image in JPG format (JPG bytes)
- `base64_image`: Base64 encoded image string
- `dataset_name`: Source dataset name

## Processing

- Images resized to max 800px on longest side (preserving aspect ratio)
- Converted to JPEG format
- Quality: 90%
- Images smaller than 800px on longest side are kept at original size

## Source Datasets and Licenses

This dataset is compiled from the following sources:

1. **TextOCR Dataset** (`yunusserhat/TextOCR-Dataset`)
   - License: CC-BY-4.0
   - Source: https://huggingface.co/datasets/yunusserhat/TextOCR-Dataset

2. **Open Images v7** (`bitmind/open-images-v7`)
   - License: CC-BY-2.0
   - Source: https://huggingface.co/datasets/bitmind/open-images-v7

3. **MPII Human Pose Dataset** (`Voxel51/MPII_Human_Pose_Dataset`)
   - License: BSD-2-Clause
   - Source: https://huggingface.co/datasets/Voxel51/MPII_Human_Pose_Dataset

4. **Nature Dataset** (`mertcobanov/nature-dataset`)
   - License: -
   - Source: https://huggingface.co/datasets/mertcobanov/nature-dataset

5. **Describe Anything Dataset - COCOStuff** (`nvidia/describe-anything-dataset`)
   - License: COCO dataset license
   - Source: https://huggingface.co/datasets/nvidia/describe-anything-dataset

6. **Describe Anything Dataset - LVIS** (`nvidia/describe-anything-dataset`)
   - License: LVIS dataset license
   - Source: https://huggingface.co/datasets/nvidia/describe-anything-dataset

7. **FineVision** (`HuggingFaceM4/FineVision`)
   - License: -
   - Source: https://huggingface.co/datasets/HuggingFaceM4/FineVision

8. **CreatiDesign Dataset** (`HuiZhang0812/CreatiDesign_dataset`)
   - License: Apache-2.0
   - Source: https://huggingface.co/datasets/HuiZhang0812/CreatiDesign_dataset

9. **Midjourney Detailed Prompts** (`MohamedRashad/midjourney-detailed-prompts`)
   - License: Apache-2.0
   - Source: https://huggingface.co/datasets/MohamedRashad/midjourney-detailed-prompts

10. **WPI Historical Image Collection** (`afrenkai/WPI-Historical-Image-Collection`)
    - License: -
    - Source: https://huggingface.co/datasets/afrenkai/WPI-Historical-Image-Collection

11. **NSFW Dataset** (`amaye15/NSFW`)
    - License: -
    - Source: https://huggingface.co/datasets/amaye15/NSFW
    - **Warning**: Contains adult content

## Intended Use

This dataset is intended **only for testing internal mechanisms** of the Media Vault Vector Database system. It is designed for:
- Validation of vector database indexing and retrieval systems
- Testing image processing pipelines
- Performance benchmarking
- Internal development and testing purposes

**This dataset is not intended for:**
- Commercial use without proper licensing verification
- Training production models without reviewing source licenses
- Public redistribution without attribution

## Usage

```python
from datasets import load_dataset

# Load specific split
sfw = load_dataset("SamoXXX/MV-VDB-photos-small", split="sfw")
x_nsfw = load_dataset("SamoXXX/MV-VDB-photos-small", split="x_nsfw")

# Access images
for example in sfw:
    index = example['index']
    image = example['image']  # PIL Image object
    base64_image = example['base64_image']  # Base64 string
    source_dataset = example['dataset_name']
    
    # Image is already a PIL Image, ready to use
    image.show()  # Display image
```

## Warning

The `x_nsfw` split contains adult content. Use responsibly and in accordance with applicable laws and regulations. This content is included for testing purposes only.

## License

This dataset compilation is released under **unknown** license. Users are responsible for:
- Verifying individual source dataset licenses
- Complying with all applicable terms and conditions
- Ensuring proper attribution when using source datasets

## Citation

If you use this dataset, please cite the original source datasets listed above. This compilation is for testing purposes only.

## Disclaimer

This dataset is provided "as-is" for internal testing purposes. The creators make no warranties about the dataset's fitness for any particular purpose. Users are responsible for ensuring compliance with all applicable licenses and regulations.