def data_generator(batch_size):
(X_train, Y_train), (X_test, Y_test) = keras.datasets.cifar10.load_data()
X_train = X_train.astype(np.float32) / 255
X_test = X_test.astype(np.float32) / 255
datagen = ImageDataGenerator(featurewise_center=True, zca_whitening=True)
datagen.fit(X_train)
X_train = datagen.standardize(X_train)
X_test = datagen.standardize(X_test)
return tf.data.Dataset.from_tensor_slices((X_train, Y_train)).batch(batch_size), \
tf.data.Dataset.from_tensor_slices((X_test, Y_test)).batch(batch_size)
me = []
se = []
ret, _ = cap.read()
cap.set(0, 0)
res = []
t = time()
i = 0
while ret:
print("Frame No: ", i)
ret, I = cap.read()
if not ret:
break
I = cv2.resize(I, (224, 224))
img = np.reshape(I, [1, 224, 224, 3])
img = img.astype('float32')
img_data = test_datagen.standardize(img)
me.append(img_data.mean())
se.append(img_data.std())
f_all = custom_resnet_model.predict(img_data)
f_all = ["%.5f" % (100 * f) for f in list(f_all[0])]
res.append(f_all)
i = i + 1
run_time = time() - t
res = np.array(res)
np.save('predictions.npy', res, allow_pickle=False)
print('Model Load Time = ' + str(model_load_time))
print('Run Time = ' + str(run_time))
class PredictGenerator(keras.utils.Sequence):
def __init__(self,
slide_path: str,
batch_size: int,
area_to_predict: Area_box = None,
tqdm_enabled=True):
self.batch_size = batch_size
self.slide = Slide(slide_path)
self.tqdm_enabled = tqdm_enabled
self.tqdm = None
self.area_to_predict = area_to_predict if area_to_predict else (
0, 0, self.slide.slide.dimensions[0],
self.slide.slide.dimensions[1])
self.datagen_for_standartize = ImageDataGenerator(
samplewise_center=True,
samplewise_std_normalization=True,
)
x = range(self.area_to_predict[0], self.area_to_predict[2] - TILE_SIZE,
TILE_STEP)
y = range(self.area_to_predict[1], self.area_to_predict[3] - TILE_SIZE,
TILE_STEP)
self.coordinates_grid = np.stack(np.meshgrid(x, y), axis=2)
self.all_coordinates = self.coordinates_grid.reshape(-1, 2)
self.label_names_to_id = DatasetPreparation.get_label_name_to_label_id_dict(
)
def __len__(self):
return int(np.ceil(len(self.all_coordinates) / self.batch_size))
def __getitem__(self, index):
if self.tqdm_enabled:
self.update_tqdm(index)
if isinstance(index, slice):
return [
self[i] for i in range(index.start, index.stop, index.step)
]
addresses = self.all_coordinates[index * self.batch_size:(index + 1) *
self.batch_size]
X = self.__data_generation(addresses)
return X, np.zeros(len(X))
def update_tqdm(self, index):
if index == 0:
self.tqdm = tqdm.tqdm(total=len(self))
self.tqdm.update(1)
if index == len(self) - 1:
self.tqdm.close()
def __data_generation(self, addresses: np.ndarray) -> np.ndarray:
data = np.asarray([
np.asarray(
self.slide.cut_tile(*add).resize(
NETWORK_INPUT_SHAPE[:2]).convert('RGB'))
for add in addresses
]).astype(np.float)
return self.datagen_for_standartize.standardize(data)
