def get_shuffle_cropped_data_batch(self, batch_size, resizing=None):
datas = []
shapes = []
# files_batch = random.sample(self.files, batch_size)
if self.idx + batch_size > len(self.files):
start_idx = -batch_size + len(self.files)
last_idx = len(self.files)
self.idx = 0
else:
start_idx = self.idx
last_idx = self.idx + batch_size
self.idx += batch_size
for file in self.files[start_idx:last_idx]:
data, shape = processing.load_image(file, meta=True)
if resizing is not None:
if resizing > min(shape[0], shape[1]):
return self.get_data_batch(batch_size, resizing)
rand_w = random.randint(0, shape[0] - resizing)
rand_h = random.randint(0, shape[1] - resizing)
data = data[rand_w:rand_w + resizing,
rand_h:rand_h + resizing, :]
# data = processing.preprocess_image(data)
datas.append(data)
shapes.append((resizing, resizing))
return (datas, shapes)
def get_datasets(self, resizing=None):
datas = []
shapes = []
for file in self.files:
data, shape = processing.load_image(file, meta=True)
if resizing is not None:
data = processing.imresize(data, (resizing, resizing))
datas.append(data)
shapes.append(shape)
return (datas, shapes)
def get_rand_data_batch(self, batch_size, resizing=None):
datas = []
shapes = []
files_batch = random.sample(self.files, batch_size)
for file in files_batch:
data, shape = processing.load_image(file, meta=True)
if resizing is not None:
data = processing.imresize(data, (resizing, resizing))
#data = processing.preprocess_image(data)
datas.append(data)
shapes.append(shape)
return (datas, shapes, files_batch)
def get_datasets(dir_name, resizing=None):
files = [
dir_name + '/' + png for png in os.listdir(dir_name)
if not os.path.isdir(dir_name + '/' + png)
]
print(files)
datas = []
shapes = []
for file in files:
data, shape = processing.load_image(file, meta=True)
if resizing is not None:
data = processing.imresize(data, (resizing, resizing),
preserve_range=True)
datas.append(data)
shapes.append(shape)
return (datas, shapes)
def get_data_batch(self, batch_size, resizing=None):
datas = []
shapes = []
#files_batch = random.sample(self.files, batch_size)
if self.idx + batch_size > len(self.files):
start_idx = -batch_size + len(self.files)
last_idx = len(self.files)
self.idx = 0
else:
start_idx = self.idx
last_idx = self.idx + batch_size
self.idx += batch_size
for file in self.files[start_idx:last_idx]:
data, shape = processing.load_image(file, meta=True)
if resizing is not None:
data = processing.imresize(data, (resizing, resizing))
# data = processing.preprocess_image(data)
datas.append(data)
shapes.append(shape)
return (datas, shapes)
import os
os.chdir(os.path.join(os.getcwd(), '../'))
print(os.getcwd())
from util import processing as util
import matplotlib.pyplot as plt
img = util.load_image('img/test.jpg', size=128)
#plt.plot(img)
img = util.preprocess_image(img)[None]
#print(img)
#print('min_val : {} . max_val : {}'.format(img.min(), img.max()))
#print('finish')