def test_img_transforms(self):
x = np.random.random((3, 200, 200))
_ = preprocessing_image.random_rotation(x, 20)
_ = preprocessing_image.random_shift(x, 0.2, 0.2)
_ = preprocessing_image.random_shear(x, 2.)
_ = preprocessing_image.random_zoom(x, (0.5, 0.5))
_ = preprocessing_image.apply_channel_shift(x, 2, 2)
_ = preprocessing_image.apply_affine_transform(x, 2)
with self.assertRaises(ValueError):
preprocessing_image.random_zoom(x, (0, 0, 0))
_ = preprocessing_image.random_channel_shift(x, 2.)
def paint_text(text,
w,
h,
rotate=True,
ud=True,
multi_fonts=True,
save_path=None):
surface = Image.new('L', (w, h), 255) # "L" represents gray
border_w_h = (2, 2)
box_width = 0
box_height = 0
for font_size in range(32, 20, -1):
if multi_fonts:
fonts = [
"simsun.ttc", "simhei.ttf", "msyh.ttc", "msyhbd.ttc",
"msyhl.ttc", "simfang.ttf", "simkai.ttf"
]
font = ImageFont.truetype(np.random.choice(fonts), font_size)
else:
font = ImageFont.truetype("msyh.ttc", font_size)
box_width, box_height = font.getsize(text)
if box_width <= (w - 2 * border_w_h[0]) and box_height <= (
h - border_w_h[1]):
break
elif font_size <= 21:
raise IOError(
('Could not fit string into image.'
'Max char count is too large for given image width.'))
max_shift_x = w - box_width - border_w_h[0]
max_shift_y = h - box_height - border_w_h[1]
top_left_x = np.random.randint(border_w_h[0], int(max_shift_x) + 1)
if ud:
top_left_y = np.random.randint(0, int(max_shift_y) + 1)
else:
top_left_y = max_shift_y // 2
draw = ImageDraw.Draw(surface)
draw.text((top_left_x, top_left_y), text, fill=0, font=font)
a = np.array(surface, np.uint8)
a = a.astype(np.float32) / 255
a = np.expand_dims(a, 0)
if rotate:
a = image.random_rotation(
a, 70 * (np.min([top_left_y, h - box_height - top_left_y])) / w)
a = speckle(a)
if save_path:
save_array = a[0, :, :] * 255
save_array = save_array.astype(np.uint8)
save_image = Image.fromarray(save_array)
save_image.save(save_path)
return a
def paint_text(text, w, h, rotate=False, ud=False, multi_fonts=False):
surface = Image.new('L', (w, h), 255) # "L" represents gray
border_w_h = (2, 2)
box_width = 0
box_height = 0
for font_size in range(32, 20, -1):
if multi_fonts:
fonts = [
"simsun.ttc", "simhei.ttf", "msyh.ttc", "msyhbd.ttc",
"msyhl.ttc", "simfang.ttf", "simkai.ttf"
]
font = ImageFont.truetype(np.random.choice(fonts), font_size)
else:
font = ImageFont.truetype("msyh.ttc", font_size)
box_width, box_height = font.getsize(text)
if box_width <= (w - 2 * border_w_h[0]) and box_height <= (
h - border_w_h[1]):
break
elif font_size <= 21:
raise IOError(
('Could not fit string into image.'
'Max char count is too large for given image width.'))
# teach the RNN translational invariance by
# fitting text box randomly on canvas, with some room to rotate
max_shift_x = w - box_width - border_w_h[0]
max_shift_y = h - box_height - border_w_h[1]
top_left_x = np.random.randint(border_w_h[0], int(max_shift_x) + 1)
if ud:
top_left_y = np.random.randint(0, int(max_shift_y) + 1)
else:
top_left_y = max_shift_y // 2
draw = ImageDraw.Draw(surface)
draw.text((top_left_x, top_left_y), text, fill=0, font=font)
a = np.array(surface, np.uint8)
a = a.astype(np.float32) / 255
a = np.expand_dims(a, 0)
if rotate: # avoid string exceeding canvas
a = image.random_rotation(
a, 70 * (np.min([top_left_y, h - box_height - top_left_y])) / w)
a = speckle(a)
return a