Dataset Viewer
Auto-converted to Parquet Duplicate
image
image
text
string
source
string
image_id
string
has_label
bool
null
book32
image-000000001
false
null
book32
image-000000002
false
null
book32
image-000000003
false
null
book32
image-000000004
false
null
book32
image-000000005
false
null
book32
image-000000006
false
null
book32
image-000000007
false
null
book32
image-000000008
false
null
book32
image-000000009
false
null
book32
image-000000010
false
null
book32
image-000000011
false
null
book32
image-000000012
false
null
book32
image-000000013
false
null
book32
image-000000014
false
null
book32
image-000000015
false
null
book32
image-000000016
false
null
book32
image-000000017
false
null
book32
image-000000018
false
null
book32
image-000000019
false
null
book32
image-000000020
false
null
book32
image-000000021
false
null
book32
image-000000022
false
null
book32
image-000000023
false
null
book32
image-000000024
false
null
book32
image-000000025
false
null
book32
image-000000026
false
null
book32
image-000000027
false
null
book32
image-000000028
false
null
book32
image-000000029
false
null
book32
image-000000030
false
null
book32
image-000000031
false
null
book32
image-000000032
false
null
book32
image-000000033
false
null
book32
image-000000034
false
null
book32
image-000000035
false
null
book32
image-000000036
false
null
book32
image-000000037
false
null
book32
image-000000038
false
null
book32
image-000000039
false
null
book32
image-000000040
false
null
book32
image-000000041
false
null
book32
image-000000042
false
null
book32
image-000000043
false
null
book32
image-000000044
false
null
book32
image-000000045
false
null
book32
image-000000046
false
null
book32
image-000000047
false
null
book32
image-000000048
false
null
book32
image-000000049
false
null
book32
image-000000050
false
null
book32
image-000000051
false
null
book32
image-000000052
false
null
book32
image-000000053
false
null
book32
image-000000054
false
null
book32
image-000000055
false
null
book32
image-000000056
false
null
book32
image-000000057
false
null
book32
image-000000058
false
null
book32
image-000000059
false
null
book32
image-000000060
false
null
book32
image-000000061
false
null
book32
image-000000062
false
null
book32
image-000000063
false
null
book32
image-000000064
false
null
book32
image-000000065
false
null
book32
image-000000066
false
null
book32
image-000000067
false
null
book32
image-000000068
false
null
book32
image-000000069
false
null
book32
image-000000070
false
null
book32
image-000000071
false
null
book32
image-000000072
false
null
book32
image-000000073
false
null
book32
image-000000074
false
null
book32
image-000000075
false
null
book32
image-000000076
false
null
book32
image-000000077
false
null
book32
image-000000078
false
null
book32
image-000000079
false
null
book32
image-000000080
false
null
book32
image-000000081
false
null
book32
image-000000082
false
null
book32
image-000000083
false
null
book32
image-000000084
false
null
book32
image-000000085
false
null
book32
image-000000086
false
null
book32
image-000000087
false
null
book32
image-000000088
false
null
book32
image-000000089
false
null
book32
image-000000090
false
null
book32
image-000000091
false
null
book32
image-000000092
false
null
book32
image-000000093
false
null
book32
image-000000094
false
null
book32
image-000000095
false
null
book32
image-000000096
false
null
book32
image-000000097
false
null
book32
image-000000098
false
null
book32
image-000000099
false
null
book32
image-000000100
false
End of preview. Expand in Data Studio

YAML Metadata Warning:The task_categories "text-recognition" is not in the official list: text-classification, token-classification, table-question-answering, question-answering, zero-shot-classification, translation, summarization, feature-extraction, text-generation, fill-mask, sentence-similarity, text-to-speech, text-to-audio, automatic-speech-recognition, audio-to-audio, audio-classification, audio-text-to-text, voice-activity-detection, depth-estimation, image-classification, object-detection, image-segmentation, text-to-image, image-to-text, image-to-image, image-to-video, unconditional-image-generation, video-classification, reinforcement-learning, robotics, tabular-classification, tabular-regression, tabular-to-text, table-to-text, multiple-choice, text-ranking, text-retrieval, time-series-forecasting, text-to-video, image-text-to-text, image-text-to-image, image-text-to-video, visual-question-answering, document-question-answering, zero-shot-image-classification, graph-ml, mask-generation, zero-shot-object-detection, text-to-3d, image-to-3d, image-feature-extraction, video-text-to-text, keypoint-detection, visual-document-retrieval, any-to-any, video-to-video, other

YAML Metadata Warning:The task_categories "self-supervised-learning" is not in the official list: text-classification, token-classification, table-question-answering, question-answering, zero-shot-classification, translation, summarization, feature-extraction, text-generation, fill-mask, sentence-similarity, text-to-speech, text-to-audio, automatic-speech-recognition, audio-to-audio, audio-classification, audio-text-to-text, voice-activity-detection, depth-estimation, image-classification, object-detection, image-segmentation, text-to-image, image-to-text, image-to-image, image-to-video, unconditional-image-generation, video-classification, reinforcement-learning, robotics, tabular-classification, tabular-regression, tabular-to-text, table-to-text, multiple-choice, text-ranking, text-retrieval, time-series-forecasting, text-to-video, image-text-to-text, image-text-to-image, image-text-to-video, visual-question-answering, document-question-answering, zero-shot-image-classification, graph-ml, mask-generation, zero-shot-object-detection, text-to-3d, image-to-3d, image-feature-extraction, video-text-to-text, keypoint-detection, visual-document-retrieval, any-to-any, video-to-video, other

Union14M-U-STR: Unlabeled Scene Text Recognition Dataset

Dataset Description

Union14M-U-STR contains 10M unlabeled images collected from 3 large datasets for self-supervised learning in Scene Text Recognition (STR). Images are collected using IoU Voting method.

Key Features

  • 10M unlabeled images for self-supervised learning
  • 3 data sources: book32, common_crawl, openvino
  • No text labels - designed for self-supervised approaches
  • IoU Voting collection method

Dataset Structure

{
    "image": PIL.Image,
    "text": null,
    "source": str,
    "image_id": str,
    "has_label": false
}

Splits

  • train: All unlabeled images (no validation/test splits for unlabeled data)

Data Sources

  • book32: Book-related text images
  • common_crawl: Web images from Common Crawl
  • openvino: Images from OpenVINO dataset

Usage

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("Bekhouche/Union14M-U-STR")

# Access training data
train_data = dataset["train"]

# Example usage for self-supervised learning
for sample in train_data:
    image = sample["image"]
    source = sample["source"]
    # Use for self-supervised STR training...

Citation

If you use this dataset, please cite the original Union14M paper.

Downloads last month
18