def test_recognize_training():
c.clear_temp()
seq = 'JWP26'
# seq = 'K464J'
# Sticking together
# seq = 'WMQPQ'
# seq = '14FWX'
# seq = '4TJ3R'
# seq = '5PW9Y'
# seq = '6ML6X'
# seq = '48HXH'
# seq = 'Y581K'
# Isolation
# seq = 'QN4EL'
# Complicated
# seq = '2XML9'
# seq = 'W9WU4'
if seq:
image = dataset_manager.get_training_image(seq)
else:
seq, image = dataset_manager.get_training_image()
success, seq_r, weak_confidence = CaptchaRecognizer().recognize(
image, verbose=True, save_intermediate=True, force_partition=True)
if success:
if weak_confidence:
print('Weak confidence')
print('Recognized is', seq_r)
print('Actual is', seq)
print('Result: {}'.format(seq == seq_r))
def test_captcha_recognition():
c.clear_temp()
seq = None
# Below are all the training images that are partitioned falsely
# seq = 'YFF5M'
# seq = 'W1PM4'
# seq = 'W1R4R'
# seq = 'YTM6X'
# seq = 'W9WU4'
# seq = 'EFTWY'
# seq = '5WTGP'
# seq = '113W2'
# seq = 'UWFG1'
# Sticking together
# seq = 'WMQPQ'
# seq = '14FWX'
# seq = '4TJ3R'
# seq = '5PW9Y'
# seq = '6ML6X'
# seq = '48HXH'
# seq = 'Y581K'
# Isolation
# seq = 'QN4EL'
# Complicated
# seq = '2XML9'
# seq = 'W9WU4'
# image = dataset_manager.get_training_image(seq)
# image = dataset_manager.get_test_image(seq)
a = BilibiliCaptchaProvider()
image = a.fetch()
success, captcha = CaptchaRecognizer().recognize(image,
save_intermediate=True,
verbose=True,
reoptimize=False)
if success:
print(captcha)
print('Recognized captcha is ',
BilibiliCaptchaProvider().verify(captcha))
def tune_partition_parameter():
h_tol = np.arange(4, 8)
s_tol = np.arange(30, 50, 4)
v_tol = np.arange(50, 70, 4)
rate = np.zeros((len(h_tol), len(s_tol), len(v_tol)))
for h in h_tol:
for s in s_tol:
for v in v_tol:
recognizer = CaptchaRecognizer(_captcha_provider, h / 360,
s / 100, v / 100)
rate[
h - 4, (s - 30)/4, (v - 50)/4] = \
partition_training_images_to_chars(recognizer,
force_update=True,
save=False)
print(np.unravel_index(rate.argmax(), rate.shape))
print(rate.max())
np.save(os.path.join(_dataset_dir, 'gridsearch.npy'), rate)
return rate
def partition_training_images_to_chars(
captcha_recognizer=CaptchaRecognizer(), force_update=False, save=True):
time_start = time.time()
try:
json_dict = json.load(open(_PARTITION_JSON))
except Exception as e:
print(e)
print('Warning: failed to load {}. Reconstructing...'.format(
_PARTITION_JSON))
json_dict = {}
force_update = True
if force_update:
json_dict[_FAIL] = []
json_dict[_SUCCESS] = []
for char in _captcha_provider.chars:
json_dict[_NUM_CHAR.format(char)] = [0, 0]
seqs = _list_basename(_training_set_dir)
num_total = len(seqs)
old_seq_set = set(json_dict[_FAIL] + json_dict[_SUCCESS])
def seq_filter(s):
return s not in old_seq_set
seqs = list(filter(seq_filter, seqs))
num_update = len(seqs)
num_update_success = 0
recognizer = captcha_recognizer
for n in range(num_update):
seq = seqs[n]
if save:
print('{}/{}: {}'.format(n, num_update, seq))
img = get_training_image(seq)
char_images, _ = recognizer.partition(img, force_partition=False)
# If successful
if char_images is not None:
json_dict[_SUCCESS].append(seq)
num_update_success += 1
for i in range(_captcha_provider.seq_length):
char = seq[i]
json_dict[_NUM_CHAR.format(char)][1] += 1
json_dict[_NUM_CHAR.format(char)][0] += 1
if save:
path = _get_training_char_path(
char, _add_suffix('{}.{}'.format(seq, i + 1)))
mpimg.imsave(path, char_images[i], cmap=_cm_greys)
else:
json_dict[_FAIL].append(seq)
for i in range(_captcha_provider.seq_length):
char = seq[i]
json_dict[_NUM_CHAR.format(char)][1] += 1
num_total_success = len(json_dict[_SUCCESS])
json_dict[_NUM_TOTAL] = num_total
json_dict[_NUM_FAIL] = num_total - num_total_success
json_dict[_NUM_SUCCESS] = num_total_success
total_success_rate = num_total_success / num_total if num_total else 0
json_dict[_SUCCESS_RATE] = '{:.3%}'.format(total_success_rate)
json_dict[_FAIL].sort()
json_dict[_SUCCESS].sort()
json.dump(json_dict, open(_PARTITION_JSON, 'w'), sort_keys=True, indent=2)
if save:
print('Update: {}'.format(num_update))
print('Update success: {}'.format(num_update_success))
if num_update:
print('Update success rate is: {}'.format(num_update_success /
num_update))
print('Total: {}'.format(num_total))
print('Total success: {}'.format(num_total_success))
print('Total success rate is: {}'.format(total_success_rate))
time_end = time.time()
print('Elapsed time of partitioning training images: {}'.format(
time_end - time_start))
if not save:
print('h_tol = {}'.format(recognizer.h_tolerance))
print('s_tol = {}'.format(recognizer.s_tolerance))
print('v_tol = {}'.format(recognizer.v_tolerance))
print('Total success rate is: {}'.format(total_success_rate))
return total_success_rate
def test_recognize_http(show_img=False, num=1, reconstruct=False, force_partition=True):
time_start = time.time()
provider = BilibiliCaptchaProvider()
recognizer = CaptchaRecognizer()
fail = 0
right_strong = 0
right_weak = 0
wrong_strong = 0
wrong_weak = 0
for i in range(num):
image = time_func("fetch" if num == 1 else None, lambda: provider.fetch())
if show_img and num == 1:
show_image(image)
if num == 1:
success, seq, weak_confidence = recognizer.recognize(
image, save_intermediate=True, verbose=True, reconstruct=reconstruct, force_partition=force_partition
)
else:
if i == 0:
success, seq, weak_confidence = recognizer.recognize(
image,
save_intermediate=False,
verbose=False,
reconstruct=reconstruct,
force_partition=force_partition,
)
else:
success, seq, weak_confidence = recognizer.recognize(
image, save_intermediate=False, verbose=False, reconstruct=False, force_partition=force_partition
)
if success:
print(seq)
result = time_func("verify" if num == 1 else None, lambda: provider.verify(seq))
if num == 1:
print("Recognized seq is {}".format(result))
if result:
if weak_confidence:
right_weak += 1
else:
right_strong += 1
else:
if weak_confidence:
wrong_weak += 1
else:
wrong_strong += 1
else:
fail += 1
right_total = right_strong + right_weak
wrong_total = wrong_strong + wrong_weak
print("Fail: ", fail)
print("Right weak: ", right_weak)
print("Right strong: ", right_strong)
print("Right total: ", right_total)
print("Wrong weak: ", wrong_weak)
print("Wrong strong: ", wrong_strong)
print("Wrong total: ", wrong_total)
print("Total success rate: ", (right_weak + right_strong) / num)
print("Success rate when confident: ", (right_strong + right_weak) / (num - fail) if num - fail > 0 else 0)
print(
"Success rate when strongly confident: ",
right_strong / (right_strong + wrong_strong) if right_strong + wrong_strong > 0 else 0,
)
print(
"Success rate when weakly confident: ",
right_weak / (right_weak + wrong_weak) if right_weak + wrong_weak > 0 else 0,
)
time_end = time.time()
print("Time used to test recognize http is: ", time_end - time_start)
def test_recognize_http(show_img=False,
num=1,
reconstruct=False,
force_partition=True):
time_start = time.time()
provider = BilibiliCaptchaProvider()
recognizer = CaptchaRecognizer()
fail = 0
right_strong = 0
right_weak = 0
wrong_strong = 0
wrong_weak = 0
for i in range(num):
image = time_func('fetch' if num == 1 else None,
lambda: provider.fetch())
if show_img and num == 1:
show_image(image)
if num == 1:
success, seq, weak_confidence = recognizer.recognize(
image,
save_intermediate=True,
verbose=True,
reconstruct=reconstruct,
force_partition=force_partition)
else:
if i == 0:
success, seq, weak_confidence = recognizer.recognize(
image,
save_intermediate=False,
verbose=False,
reconstruct=reconstruct,
force_partition=force_partition)
else:
success, seq, weak_confidence = recognizer.recognize(
image,
save_intermediate=False,
verbose=False,
reconstruct=False,
force_partition=force_partition)
if success:
print(seq)
result = time_func('verify' if num == 1 else None,
lambda: provider.verify(seq))
if num == 1:
print('Recognized seq is {}'.format(result))
if result:
if weak_confidence:
right_weak += 1
else:
right_strong += 1
else:
if weak_confidence:
wrong_weak += 1
else:
wrong_strong += 1
else:
fail += 1
right_total = right_strong + right_weak
wrong_total = wrong_strong + wrong_weak
print('Fail: ', fail)
print('Right weak: ', right_weak)
print('Right strong: ', right_strong)
print('Right total: ', right_total)
print('Wrong weak: ', wrong_weak)
print('Wrong strong: ', wrong_strong)
print('Wrong total: ', wrong_total)
print('Total success rate: ', (right_weak + right_strong) / num)
print('Success rate when confident: ',
(right_strong + right_weak) / (num - fail) if num - fail > 0 else 0)
print(
'Success rate when strongly confident: ',
right_strong / (right_strong + wrong_strong)
if right_strong + wrong_strong > 0 else 0)
print(
'Success rate when weakly confident: ', right_weak /
(right_weak + wrong_weak) if right_weak + wrong_weak > 0 else 0)
time_end = time.time()
print('Time used to test recognize http is: ', time_end - time_start)