Scene text image data with text words and their coordinates annotations collected from LAION-400M dataset.
We provide the links of images from LAION-400M dataset, and our mannually annotated text and coordinates information. Currently, we release 5000 samples with 49866 text annotations with center coordinates.
The following is one sample of the annotation:
{
"index": "part1_000228186", # just a random name
"url": "https://thesiliconreview.net/administrator/article_image/siliconreview-Googles-face-recognition-technology-to-be-added-into-home-security-camera.jpg", # url of the image, we do not provide downloading following the license of LAION-400M
"width": 840, # image width
"height": 449, # image height
"caption": "Google's face recognition technology to be added into home-security camera", # caption copied from LAION-400M
"annotations": [ # the annotation of text instances within this image
{
"text": "captured",
"center": [
183.0,
57.7
]
},
{
"text": "Face",
"center": [
124.4,
53.0
]
},
{
"text": "###", # means annotators could not read the text content from the image
"center": [
117.0,
2.0
]
},
{
"text": "###", # means annotators could not read the text content from the image
"center": [
166.0,
2.0
]
}
]
},Click to download LAION-TextData dataset from Baidu Pan.
Following the LAION-400M, LAION-TextData dataset was built for research purposes, and is not meant for any real-world production or application, and is under the most open Creative Common CC-BY 4.0 license. The images are under their copyright.