class Eval():
def __init__(self):
self.img_processing = ImageProcessing()
self.src_dir = 'test_imgs'
self.imgs_coords = self.img_processing.generate_copped_imgs(
self.src_dir, 32, 32, 64, 64)
self.answer_sheet = {
'1.png': [73, 74],
'2.png': [875, 874],
'3.png': [146, 147]
}
"""
best second
# [1,33](73) [1,32](72) 1.png
# [13,43](875) [13,42](874) 3.png
# [3,20](146) [3,21](147) 2.png
"""
def check_preds(self, preds, answer_indices):
"""
:param imgs:
:param coords:
:param preds:
:param answer_indices:
:return:
"""
max_idx = np.argmax(preds, axis=0)
max_value = np.max(preds, axis=0)
# debug
print 'max idx : {} , max value : {} , answer_indicies : {}'.format(
max_idx, max_value, answer_indices)
answer = False
if max_value > 0.5:
if max_idx in answer_indices:
answer = True
else:
answer = False
return answer
def validate(self, imgs, sess_op, preds_op, batch_size, x_op, phase_train,
normalize):
batch_list = self.divide_images(imgs, batch_size)
preds_list = []
for batch_xs in batch_list:
if normalize and np.max(batch_xs) > 1:
batch_xs = batch_xs / 255.
preds = sess_op.run(preds_op, {x_op: batch_xs, phase_train: False})
preds_list.extend(preds)
assert len(preds_list) == len(imgs)
return preds_list
def get_acc(self, sess_op, preds_op, batch_size, x_op, phase_train):
answers = []
for key in self.imgs_coords.keys():
imgs, coords = self.imgs_coords[key]
answer_indices = self.answer_sheet[key]
preds = self.validate(imgs,
sess_op,
preds_op,
batch_size,
x_op,
phase_train,
normalize=True)
answer = self.check_preds(preds, answer_indices)
answers.append(answer)
return np.sum(answers) / float(len(answers))
def divide_images(self, images, batch_size):
debug_flag_lv0 = False
debug_flag_lv1 = False
if __debug__ == debug_flag_lv0:
print 'debug start | utils.py | divide_images'
batch_img_list = []
share = len(images) / batch_size
# print len(images)
# print len(labels)
# print 'share :',share
for i in range(share + 1):
if i == share:
imgs = images[i * batch_size:]
# print i+1, len(imgs), len(labs)
batch_img_list.append(imgs)
if __debug__ == debug_flag_lv1:
print "######utils.py: divide_images_from_batch debug mode#####"
print 'total :', len(
images), 'batch', i * batch_size, '-', len(images)
else:
imgs = images[i * batch_size:(i + 1) * batch_size]
# print i , len(imgs) , len(labs)
batch_img_list.append(imgs)
if __debug__ == debug_flag_lv1:
print "######utils.py: divide_images_from_batch debug mode######"
print 'total :', len(
images), 'batch', i * batch_size, ":", (i +
1) * batch_size
return batch_img_list
