# lines = [s for s in f.readlines()]
#lines = [s.strip() for s in lines if not s.startswith('#')]
with open("task/trainset.txt") as f:
training_set = set([s.strip() for s in f.readlines()])
with open("task/validationset1.txt") as f:
val1_set = set([s.strip() for s in f.readlines()])
with open("task/validationset2.txt") as f:
val2_set = set([s.strip() for s in f.readlines()])
with open("task/testset.txt") as f:
test_set = set([s.strip() for s in f.readlines()])
author_mapping, lines_gts, word_gts = prep_iam_writer_map.get_mapping('xml')
def prep_set(data_set, lines_gts, author_mapping):
all_authors = defaultdict(lambda: [])
all_authors_baseline = defaultdict(lambda: [])
for d in data_set:
split_name = d.split('-')
folder_and_base = "{}/{}-{}/{}".format(split_name[0], split_name[0], split_name[1], d)
image_file = "lines/{}.png".format(folder_and_base)
line_id = "{}-{}".format(split_name[0], split_name[1])
img = cv2.imread(image_file, 0)
if img is None:
print "There was an issue with ", image_file
continue
def get_gt(in_set):
author_mapping, lines_gts, word_gts = prep_iam_writer_map.get_mapping('xml')
data_set = prep_data_set(word_gts, in_set)
return data_set