Datasetpublic

nagohachi/japanese-str-dataset-v1

STR Dataset Japanese STR (Scene Text Recognition) dataset in WebDataset format. This dataset is composed of: Images of Japanese named entities (full names and their affiliations) Images of sentences retrieved from Aozora Bunko (青空文庫) and their corresponding ground truth texts. All images are synthesized using TRDG. Dataset Structure Split Samples Shards train 10,000,000 1000 valid 50,000 5 test 50,000 5 Total 10,100,000 1010… See the full description on the dataset page: https://huggingface.co/datasets/nagohachi/japanese-str-dataset-v1.

sourceHugging Facecc-by-4.0updated 9mo agoView on Hugging Face

0likes161downloads

Dataset Card

STR Dataset

Japanese STR (Scene Text Recognition) dataset in WebDataset format.

This dataset is composed of:

—Images of Japanese named entities (full names and their affiliations)
—Images of sentences retrieved from Aozora Bunko (青空文庫)

and their corresponding ground truth texts.

All images are synthesized using TRDG.

Dataset Structure

Split	Samples	Shards
train	10,000,000	1000
valid	50,000	5
test	50,000	5
Total	10,100,000	1010

Usage

python

import webdataset as wds

base_url = "https://huggingface.co/datasets/nagohachi/japanese-str-dataset-v1/resolve/main"

# Load train split
train_dataset = (
    wds.WebDataset(base_url + "/train/train-{00000..00999}.tar")
    .decode("pil")
    .to_tuple("png", "txt")
)

for image, text in train_dataset:
    # image: PIL Image
    # text: str
    pass

Format

Each sample contains:

—png: Image file (PNG format)
—txt: Ground truth text