def process(self, image_array):
model = model.load('my_model.h5')
digit = prep.crop_image(image_array)
digit = prep.crop_image(digit)
digit = prep.center_image(digit)
digit = prep.resize_image(digit)
digit = prep.min_max_scaler(digit)
digit = prep.reshape_array(digit)
digit = model.predict(digit)
return digit
def process(image_array):
model = load_model('my_model2.h5')
digit = prep.crop_image(image_array)
digit = prep.crop_image(digit)
digit = prep.center_image(digit)
digit = prep.resize_image(digit)
digit = prep.min_max_scaler(digit)
digit = prep.reshape_array(digit)
digit = model.predict(digit)
i = "Draw Again!"
predict = False
count = 0
for number in digit[0]:
if np.max(digit[0]) == number and number > 0.70:
print(number)
predict = True
break
else:
count = count + 1
if predict == True:
return count
else:
return i
# Plot Training Summaries
plot_summaries(history, PLOT_NAME1, PLOT_NAME2)
# Create Predictions
row_ids, targets = [], []
id = 0
# Loop through parquet files
for i in range(4):
img_df = pd.read_parquet(
os.path.join(DATA_DIR, 'test_image_data_' + str(i) + '.parquet'))
img_df = img_df.drop('image_id', axis=1)
# Loop through rows in parquet file
for index, row in img_df.iterrows():
img = resize_image(row.values, WIDTH, HEIGHT, WIDTH_NEW, HEIGHT_NEW)
img = np.stack((img, ) * CHANNELS, axis=-1)
image = img.reshape(-1, HEIGHT_NEW, WIDTH_NEW, 3)
# Predict
preds = model.predict(image, verbose=1)
for k in range(3):
row_ids.append('Test_' + str(id) + '_' + tgt_cols[k])
targets.append(np.argmax(preds[k]))
id += 1
# Create and Save Submission File
submission = pd.DataFrame({
'row_id': row_ids,
'target': targets
},
def generator(self):
while True:
batches = _make_batches(size=self.total_images,
batch_size=self.batch_size)
for start, end in batches:
arr = []
labels = []
cur_batch = self.image_paths[start:end]
for image_path in cur_batch:
# print image_path
img = imread(
fname=os.path.join(self.data_path, image_path))
# if channels are not 3
ndim = len(img.shape)
if ndim == 2:
img = img[..., np.newaxis]
img = np.tile(A=img, reps=(1, 1, 3))
if ndim == 4:
img = img[..., :3]
# resizing image maintaining aspect ratio
img = resize_image(img=img, size=self.input_size)
if self.training:
# random cropping while training
img = random_crop_image(img=img, size=self.input_size)
img = augment(img=img,
horizontal_flip=True,
vertical_flip=True,
brightness=True,
contrast=True,
rotation=True,
translation=True,
blur=True,
noise=True)
else:
# center cropping
h, w, c = img.shape
center_h = h / 2
center_w = w / 2
center_new_img = self.input_size / 2
new_x1 = center_w - center_new_img
new_y1 = center_h - center_new_img
new_x2 = center_w + center_new_img
new_y2 = center_h + center_new_img
if self.input_size % 2 == 1:
new_x2 += 1
new_y2 += 1
img = img[new_y1:new_y2, new_x1:new_x2]
arr.append(img)
cls = image_path.split('/')[0]
id_for_cls = self.cls2id[cls]
labels.append(id_for_cls)
arr = np.array(arr)
arr.astype('float32')
# making mean of data 0 with standard deviation 1
arr /= 255.
arr -= 0.5
arr *= 2.
# one hot encoding
labels = to_categorical(y=labels,
num_classes=self.total_classes)
yield (arr, labels)