Datasets:
feat: add script
Browse files
2d-masks-presentation-attack-detection.py
ADDED
|
@@ -0,0 +1,109 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import datasets
|
| 2 |
+
import pandas as pd
|
| 3 |
+
|
| 4 |
+
_CITATION = """\
|
| 5 |
+
@InProceedings{huggingface:dataset,
|
| 6 |
+
title = {selfie_and_video},
|
| 7 |
+
author = {TrainingDataPro},
|
| 8 |
+
year = {2023}
|
| 9 |
+
}
|
| 10 |
+
"""
|
| 11 |
+
|
| 12 |
+
_DESCRIPTION = """\
|
| 13 |
+
4000 people in this dataset. Each person took a selfie on a webcam,
|
| 14 |
+
took a selfie on a mobile phone. In addition, people recorded video from
|
| 15 |
+
the phone and from the webcam, on which they pronounced a given set of numbers.
|
| 16 |
+
Includes folders corresponding to people in the dataset. Each folder includes
|
| 17 |
+
8 files (4 images and 4 videos).
|
| 18 |
+
"""
|
| 19 |
+
_NAME = 'selfie_and_video'
|
| 20 |
+
|
| 21 |
+
_HOMEPAGE = f"https://huggingface.co/datasets/TrainingDataPro/{_NAME}"
|
| 22 |
+
|
| 23 |
+
_LICENSE = ""
|
| 24 |
+
|
| 25 |
+
_DATA = f"https://huggingface.co/datasets/TrainingDataPro/{_NAME}/resolve/main/data/"
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
class SelfieAndVideo(datasets.GeneratorBasedBuilder):
|
| 29 |
+
"""Small sample of image-text pairs"""
|
| 30 |
+
|
| 31 |
+
def _info(self):
|
| 32 |
+
return datasets.DatasetInfo(
|
| 33 |
+
description=_DESCRIPTION,
|
| 34 |
+
features=datasets.Features({
|
| 35 |
+
'photo_1': datasets.Image(),
|
| 36 |
+
'photo_2': datasets.Image(),
|
| 37 |
+
'video_3': datasets.Value('string'),
|
| 38 |
+
'video_4': datasets.Value('string'),
|
| 39 |
+
'photo_5': datasets.Image(),
|
| 40 |
+
'photo_6': datasets.Image(),
|
| 41 |
+
'video_7': datasets.Value('string'),
|
| 42 |
+
'video_8': datasets.Value('string'),
|
| 43 |
+
'set_id': datasets.Value('string'),
|
| 44 |
+
'worker_id': datasets.Value('string'),
|
| 45 |
+
'age': datasets.Value('int8'),
|
| 46 |
+
'country': datasets.Value('string'),
|
| 47 |
+
'gender': datasets.Value('string')
|
| 48 |
+
}),
|
| 49 |
+
supervised_keys=None,
|
| 50 |
+
homepage=_HOMEPAGE,
|
| 51 |
+
citation=_CITATION,
|
| 52 |
+
)
|
| 53 |
+
|
| 54 |
+
def _split_generators(self, dl_manager):
|
| 55 |
+
images = dl_manager.download(f"{_DATA}data.tar.gz")
|
| 56 |
+
annotations = dl_manager.download(f"{_DATA}{_NAME}.csv")
|
| 57 |
+
images = dl_manager.iter_archive(images)
|
| 58 |
+
return [
|
| 59 |
+
datasets.SplitGenerator(name=datasets.Split.TRAIN,
|
| 60 |
+
gen_kwargs={
|
| 61 |
+
"images": images,
|
| 62 |
+
'annotations': annotations
|
| 63 |
+
}),
|
| 64 |
+
]
|
| 65 |
+
|
| 66 |
+
def _generate_examples(self, images, annotations):
|
| 67 |
+
annotations_df = pd.read_csv(annotations, sep=';')
|
| 68 |
+
images_data = pd.DataFrame(columns=['Link', 'Bytes'])
|
| 69 |
+
for idx, (image_path, image) in enumerate(images):
|
| 70 |
+
if image_path.lower().endswith('.jpg'):
|
| 71 |
+
images_data.loc[idx] = {
|
| 72 |
+
'Link': image_path,
|
| 73 |
+
'Bytes': image.read()
|
| 74 |
+
}
|
| 75 |
+
|
| 76 |
+
annotations_df = pd.merge(annotations_df,
|
| 77 |
+
images_data,
|
| 78 |
+
on=['Link'],
|
| 79 |
+
how='left')
|
| 80 |
+
|
| 81 |
+
for idx, worker_id in enumerate(pd.unique(annotations_df['WorkerId'])):
|
| 82 |
+
annotation = annotations_df.loc[annotations_df['WorkerId'] ==
|
| 83 |
+
worker_id]
|
| 84 |
+
annotation = annotation.sort_values(['Link'])
|
| 85 |
+
data = {
|
| 86 |
+
(f'photo_{row[7][37]}' if row[7].lower().endswith('.jpg') else f'video_{row[7][37]}'):
|
| 87 |
+
({
|
| 88 |
+
'path': row[7],
|
| 89 |
+
'bytes': row[8]
|
| 90 |
+
} if row[7].lower().endswith('.jpg') else row[7])
|
| 91 |
+
for row in annotation.itertuples()
|
| 92 |
+
}
|
| 93 |
+
|
| 94 |
+
age = annotation.loc[annotation['Link'].str.lower().str.endswith(
|
| 95 |
+
'1.jpg')]['Age'].values[0]
|
| 96 |
+
country = annotation.loc[annotation['Link'].str.lower().str.
|
| 97 |
+
endswith('1.jpg')]['Country'].values[0]
|
| 98 |
+
gender = annotation.loc[annotation['Link'].str.lower().str.
|
| 99 |
+
endswith('1.jpg')]['Gender'].values[0]
|
| 100 |
+
set_id = annotation.loc[annotation['Link'].str.lower().str.
|
| 101 |
+
endswith('1.jpg')]['SetId'].values[0]
|
| 102 |
+
|
| 103 |
+
data['worker_id'] = worker_id
|
| 104 |
+
data['age'] = age
|
| 105 |
+
data['country'] = country
|
| 106 |
+
data['gender'] = gender
|
| 107 |
+
data['set_id'] = set_id
|
| 108 |
+
|
| 109 |
+
yield idx, data
|