def __init__(self,
data_dir,
file_list,
label_list=None,
transforms=None,
num_workers='auto',
buffer_size=100,
parallel_method='process',
shuffle=False):
super(ChangeDetDataset, self).__init__(transforms=transforms,
num_workers=num_workers,
buffer_size=buffer_size,
parallel_method=parallel_method,
shuffle=shuffle)
self.file_list = list()
self.labels = list()
self._epoch = 0
if label_list is not None:
with open(label_list, encoding=get_encoding(label_list)) as f:
for line in f:
item = line.strip()
self.labels.append(item)
with open(file_list, encoding=get_encoding(file_list)) as f:
for line in f:
items = line.strip().split()
if len(items) > 3:
raise Exception(
"A space is defined as the separator, but it exists in image or label name {}."
.format(line))
items[0] = path_normalization(items[0])
items[1] = path_normalization(items[1])
items[2] = path_normalization(items[2])
full_path_im1 = osp.join(data_dir, items[0])
full_path_im2 = osp.join(data_dir, items[1])
full_path_label = osp.join(data_dir, items[2])
if not osp.exists(full_path_im1):
raise IOError('The image file {} is not exist!'.format(
full_path_im1))
if not osp.exists(full_path_im2):
raise IOError('The image file {} is not exist!'.format(
full_path_im2))
if not osp.exists(full_path_label):
raise IOError('The image file {} is not exist!'.format(
full_path_label))
self.file_list.append(
[full_path_im1, full_path_im2, full_path_label])
self.num_samples = len(self.file_list)
logging.info("{} samples in file {}".format(len(self.file_list),
file_list))
def generate_images_field(self, json_info, image_id):
image = {}
image["height"] = json_info["size"]["height"]
image["width"] = json_info["size"]["width"]
image["id"] = image_id + 1
json_info["path"] = path_normalization(json_info["path"])
image["file_name"] = osp.split(json_info["path"])[-1]
return image
def generate_images_field(self, img_path, image_id):
image = {}
img = cv2.imread(img_path)
image["height"] = img.shape[0]
image["width"] = img.shape[1]
image["id"] = image_id + 1
img_path = path_normalization(img_path)
image["file_name"] = osp.split(img_path)[-1]
return image
def generate_images_field(self, json_info, image_file, image_id):
image = {}
image["height"] = json_info["imageHeight"]
image["width"] = json_info["imageWidth"]
image["id"] = image_id + 1
json_img_path = path_normalization(json_info["imagePath"])
json_info["imagePath"] = osp.join(
osp.split(json_img_path)[0], image_file)
image["file_name"] = osp.split(json_info["imagePath"])[-1]
return image
def __init__(self,
data_dir,
file_list,
label_list,
transforms=None,
num_workers='auto',
buffer_size=100,
parallel_method='process',
shuffle=False):
super(VOCDetection, self).__init__(transforms=transforms,
num_workers=num_workers,
buffer_size=buffer_size,
parallel_method=parallel_method,
shuffle=shuffle)
self.file_list = list()
self.labels = list()
self._epoch = 0
annotations = {}
annotations['images'] = []
annotations['categories'] = []
annotations['annotations'] = []
cname2cid = {}
label_id = 1
with open(label_list, encoding=get_encoding(label_list)) as fr:
for line in fr.readlines():
cname2cid[line.strip()] = label_id
label_id += 1
self.labels.append(line.strip())
logging.info("Starting to read file list from dataset...")
for k, v in cname2cid.items():
annotations['categories'].append({
'supercategory': 'component',
'id': v,
'name': k
})
from pycocotools.mask import decode
ct = 0
ann_ct = 0
with open(file_list, encoding=get_encoding(file_list)) as f:
for line in f:
img_file, json_file = [osp.join(data_dir, x) \
for x in line.strip().split()[:2]]
img_file = path_normalization(img_file)
json_file = path_normalization(json_file)
if not is_pic(img_file):
continue
if not osp.isfile(json_file):
continue
if not osp.exists(img_file):
raise IOError(
'The image file {} is not exist!'.format(img_file))
with open(json_file, mode='r', \
encoding=get_encoding(json_file)) as j:
json_info = json.load(j)
im_id = np.array([ct])
im = cv2.imread(img_file)
im_w = im.shape[1]
im_h = im.shape[0]
objs = json_info['labels']
gt_bbox = np.zeros((len(objs), 4), dtype=np.float32)
gt_class = np.zeros((len(objs), 1), dtype=np.int32)
gt_score = np.ones((len(objs), 1), dtype=np.float32)
is_crowd = np.zeros((len(objs), 1), dtype=np.int32)
difficult = np.zeros((len(objs), 1), dtype=np.int32)
gt_poly = [None] * len(objs)
for i, obj in enumerate(objs):
cname = obj['name']
gt_class[i][0] = cname2cid[cname]
x1 = max(0, obj['x1'])
y1 = max(0, obj['y1'])
x2 = min(im_w - 1, obj['x2'])
y2 = min(im_h - 1, obj['y2'])
gt_bbox[i] = [x1, y1, x2, y2]
is_crowd[i][0] = 0
if 'mask' in obj:
mask_dict = {}
mask_dict['size'] = [im_h, im_w]
mask_dict['counts'] = obj['mask'].encode()
mask = decode(mask_dict)
gt_poly[i] = self.mask2polygon(mask)
annotations['annotations'].append({
'iscrowd':
0,
'image_id':
int(im_id[0]),
'bbox': [x1, y1, x2 - x1 + 1, y2 - y1 + 1],
'area':
float((x2 - x1 + 1) * (y2 - y1 + 1)),
'segmentation': [[x1, y1, x1, y2, x2, y2, x2, y1]]
if gt_poly[i] is None else gt_poly[i],
'category_id':
cname2cid[cname],
'id':
ann_ct,
'difficult':
0
})
ann_ct += 1
im_info = {
'im_id': im_id,
'image_shape': np.array([im_h, im_w]).astype('int32'),
}
label_info = {
'is_crowd': is_crowd,
'gt_class': gt_class,
'gt_bbox': gt_bbox,
'gt_score': gt_score,
'difficult': difficult
}
if None not in gt_poly:
label_info['gt_poly'] = gt_poly
voc_rec = (im_info, label_info)
if len(objs) != 0:
self.file_list.append([img_file, voc_rec])
ct += 1
annotations['images'].append({
'height':
im_h,
'width':
im_w,
'id':
int(im_id[0]),
'file_name':
osp.split(img_file)[1]
})
if not len(self.file_list) > 0:
raise Exception('not found any voc record in %s' % (file_list))
logging.info("{} samples in file {}".format(len(self.file_list),
file_list))
self.num_samples = len(self.file_list)
# matplotlib.use() must be called *before* pylab, matplotlib.pyplot,
# or matplotlib.backends is imported for the first time
# pycocotools import matplotlib
import matplotlib
matplotlib.use('Agg')
from pycocotools.coco import COCO
self.coco_gt = COCO()
self.coco_gt.dataset = annotations
self.coco_gt.createIndex()