Spaces:
Sleeping
Sleeping
File size: 6,152 Bytes
9bf4bd7 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 |
# Copyright (c) OpenMMLab. All rights reserved.
import argparse
import math
import os
import os.path as osp
import xml.etree.ElementTree as ET
import mmcv
import mmengine
from mmocr.utils import dump_ocr_data
def collect_files(img_dir, gt_dir, ratio):
"""Collect all images and their corresponding groundtruth files.
Args:
img_dir (str): The image directory
gt_dir (str): The groundtruth directory
ratio (float): Split ratio for val set
Returns:
files (list): The list of tuples (img_file, groundtruth_file)
"""
assert isinstance(img_dir, str)
assert img_dir
assert isinstance(gt_dir, str)
assert gt_dir
assert isinstance(ratio, float)
assert ratio < 1.0, 'val_ratio should be a float between 0.0 to 1.0'
ann_list, imgs_list = [], []
for img_file in os.listdir(img_dir):
ann_list.append(osp.join(gt_dir, img_file.split('.')[0] + '.xml'))
imgs_list.append(osp.join(img_dir, img_file))
all_files = list(zip(sorted(imgs_list), sorted(ann_list)))
assert len(all_files), f'No images found in {img_dir}'
print(f'Loaded {len(all_files)} images from {img_dir}')
trn_files, val_files = [], []
if ratio > 0:
for i, file in enumerate(all_files):
if i % math.floor(1 / ratio):
trn_files.append(file)
else:
val_files.append(file)
else:
trn_files, val_files = all_files, []
print(f'training #{len(trn_files)}, val #{len(val_files)}')
return trn_files, val_files
def collect_annotations(files, nproc=1):
"""Collect the annotation information.
Args:
files (list): The list of tuples (image_file, groundtruth_file)
nproc (int): The number of process to collect annotations
Returns:
images (list): The list of image information dicts
"""
assert isinstance(files, list)
assert isinstance(nproc, int)
if nproc > 1:
images = mmengine.track_parallel_progress(
load_img_info, files, nproc=nproc)
else:
images = mmengine.track_progress(load_img_info, files)
return images
def load_img_info(files):
"""Load the information of one image.
Args:
files (tuple): The tuple of (img_file, groundtruth_file)
Returns:
img_info (dict): The dict of the img and annotation information
"""
assert isinstance(files, tuple)
img_file, gt_file = files
assert osp.basename(gt_file).split('.')[0] == osp.basename(img_file).split(
'.')[0]
# read imgs while ignoring orientations
img = mmcv.imread(img_file, 'unchanged')
img_info = dict(
file_name=osp.join(osp.basename(img_file)),
height=img.shape[0],
width=img.shape[1],
segm_file=osp.join(osp.basename(gt_file)))
if osp.splitext(gt_file)[1] == '.xml':
img_info = load_xml_info(gt_file, img_info)
else:
raise NotImplementedError
return img_info
def load_xml_info(gt_file, img_info):
"""Collect the annotation information.
Annotation Format
<image>
<imageName>DSC02306.JPG</imageName>
<resolution x="640" y="480" />
<words>
<word x="61" y="140" width="566" height="107">
<character x="61" y="147" width="75" height="94" char="C" />
<character x="173" y="147" width="77" height="93" char="L" />
<character x="251" y="146" width="83" height="96" char="A" />
<character x="335" y="146" width="75" height="97" char="V" />
<character x="409" y="140" width="52" height="105" char="I" />
<character x="464" y="147" width="76" height="96" char="T" />
<character x="538" y="154" width="89" height="93" char="A" />
</word>
</words>
<illumination>no</illumination>
<difficulty>2</difficulty>
<tag>
</tag>
</image>
Args:
gt_file (str): The path to ground-truth
img_info (dict): The dict of the img and annotation information
Returns:
img_info (dict): The dict of the img and annotation information
"""
obj = ET.parse(gt_file)
root = obj.getroot()
anno_info = []
for word in root.iter('word'):
x, y = max(0, int(word.attrib['x'])), max(0, int(word.attrib['y']))
w, h = int(word.attrib['width']), int(word.attrib['height'])
bbox = [x, y, w, h]
segmentation = [x, y, x + w, y, x + w, y + h, x, y + h]
anno = dict(
iscrowd=0,
category_id=1,
bbox=bbox,
area=w * h,
segmentation=[segmentation])
anno_info.append(anno)
img_info.update(anno_info=anno_info)
return img_info
def parse_args():
parser = argparse.ArgumentParser(
description='Generate training and val set of KAIST ')
parser.add_argument('root_path', help='Root dir path of KAIST')
parser.add_argument(
'--val-ratio', help='Split ratio for val set', default=0.0, type=float)
parser.add_argument(
'--nproc', default=1, type=int, help='Number of process')
args = parser.parse_args()
return args
def main():
args = parse_args()
root_path = args.root_path
ratio = args.val_ratio
trn_files, val_files = collect_files(
osp.join(root_path, 'imgs'), osp.join(root_path, 'annotations'), ratio)
# Train set
trn_infos = collect_annotations(trn_files, nproc=args.nproc)
with mmengine.Timer(
print_tmpl='It takes {}s to convert KAIST Training annotation'):
dump_ocr_data(trn_infos, osp.join(root_path,
'instances_training.json'),
'textdet')
# Val set
if len(val_files) > 0:
val_infos = collect_annotations(val_files, nproc=args.nproc)
with mmengine.Timer(
print_tmpl='It takes {}s to convert KAIST Val annotation'):
dump_ocr_data(val_infos, osp.join(root_path, 'instances_val.json'),
'textdet')
if __name__ == '__main__':
main()
|