def load_image_array_flowers(image_file, image_size):
img = skimage.io.imread(image_file)
# GRAYSCALE
if len(img.shape) == 2:
img_new = np.ndarray( (img.shape[0], img.shape[1], 3), dtype = 'uint8')
img_new[:,:,0] = img
img_new[:,:,1] = img
img_new[:,:,2] = img
img = img_new
img_resized = skimage.transform.resize(img, (image_size, image_size))
# FLIP HORIZONTAL WIRH A PROBABILITY 0.5
if random.random() > 0.5:
img_resized = np.fliplr(img_resized)
return img_resized.astype('float32')
python类transform()的实例源码
def load_image(path):
# load image
img = skimage.io.imread(path)
img = img / 255.0
assert (0 <= img).all() and (img <= 1.0).all()
# print "Original Image Shape: ", img.shape
# we crop image from center
short_edge = min(img.shape[:2])
yy = int((img.shape[0] - short_edge) / 2)
xx = int((img.shape[1] - short_edge) / 2)
crop_img = img[yy: yy + short_edge, xx: xx + short_edge]
# resize to 224, 224
resized_img = skimage.transform.resize(crop_img, (224, 224))
return resized_img
# returns the top1 string
def load_image2(path, height=None, width=None):
# load image
img = skimage.io.imread(path)
img = img / 255.0
if height is not None and width is not None:
ny = height
nx = width
elif height is not None:
ny = height
nx = img.shape[1] * ny / img.shape[0]
elif width is not None:
nx = width
ny = img.shape[0] * nx / img.shape[1]
else:
ny = img.shape[0]
nx = img.shape[1]
return skimage.transform.resize(img, (ny, nx))
def load_augment(fname, w, h, aug_params=no_augmentation_params,
transform=None, sigma=0.0, color_vec=None):
"""Load augmented image with output shape (w, h).
Default arguments return non augmented image of shape (w, h).
To apply a fixed transform (color augmentation) specify transform
(color_vec).
To generate a random augmentation specify aug_params and sigma.
"""
img = load_image(fname)
img = perturb(img, augmentation_params=aug_params, target_shape=(w, h))
#if transform is None:
# img = perturb(img, augmentation_params=aug_params, target_shape=(w, h))
#else:
# img = perturb_fixed(img, tform_augment=transform, target_shape=(w, h))
#randString = str(np.random.normal(0,1,1))
#im = Image.fromarray(img.transpose(1,2,0).astype('uint8'))
#figName = fname.split("/")[-1]
#im.save("imgs/"+figName+randString+".jpg")
np.subtract(img, MEAN[:, np.newaxis, np.newaxis], out=img)
#np.divide(img, STD[:, np.newaxis, np.newaxis], out=img)
#img = augment_color(img, sigma=sigma, color_vec=color_vec)
return img
def load_augment(fname, w, h, aug_params=no_augmentation_params,
transform=None, sigma=0.0, color_vec=None):
"""Load augmented image with output shape (w, h).
Default arguments return non augmented image of shape (w, h).
To apply a fixed transform (color augmentation) specify transform
(color_vec).
To generate a random augmentation specify aug_params and sigma.
"""
img = load_image(fname)
if transform is None:
img = perturb(img, augmentation_params=aug_params, target_shape=(w, h))
else:
img = perturb_fixed(img, tform_augment=transform, target_shape=(w, h))
np.subtract(img, MEAN[:, np.newaxis, np.newaxis], out=img)
np.divide(img, STD[:, np.newaxis, np.newaxis], out=img)
img = augment_color(img, sigma=sigma, color_vec=color_vec)
return img
def build_centering_transform(image_shape, target_shape):
"""Image cetering transform
Args:
image_shape: tuple(rows, cols), input image shape
target_shape: tuple(rows, cols), output image shape
Returns:
a centering transform instance
"""
rows, cols = image_shape
trows, tcols = target_shape
shift_x = (cols - tcols) / 2.0
shift_y = (rows - trows) / 2.0
return skimage.transform.SimilarityTransform(translation=(shift_x, shift_y))
def build_center_uncenter_transforms(image_shape):
"""Center Unceter transform
These are used to ensure that zooming and rotation happens around the center of the image.
Use these transforms to center and uncenter the image around such a transform.
Args:
image_shape: tuple(rows, cols), input image shape
Returns:
a center and an uncenter transform instance
"""
center_shift = np.array(
[image_shape[1], image_shape[0]]) / 2.0 - 0.5 # need to swap rows and cols here apparently! confusing!
tform_uncenter = skimage.transform.SimilarityTransform(
translation=-center_shift)
tform_center = skimage.transform.SimilarityTransform(
translation=center_shift)
return tform_center, tform_uncenter
def perturb_fixed(img, tform_augment, target_shape=(50, 50), mode='constant', mode_cval=0):
"""Perturb image Determinastic
It perturbs an image with augmentation transform with determinastic params
used for validation/testing data
Args:
img: a `ndarray`, input image
augmentation_paras: a dict, with augmentation name as keys and values as params
target_shape: a tuple(rows, cols), output image shape
mode: mode for transformation
available modes: {`constant`, `edge`, `symmetric`, `reflect`, `wrap`}
mode_cval: float, Used in conjunction with mode `constant`,
the value outside the image boundaries
Returns:
a `ndarray` of transformed image
"""
shape = img.shape[1:]
tform_centering = build_centering_transform(shape, target_shape)
tform_center, tform_uncenter = build_center_uncenter_transforms(shape)
# shift to center, augment, shift back (for the rotation/shearing)
tform_augment = tform_uncenter + tform_augment + tform_center
return fast_warp(img, tform_centering + tform_augment,
output_shape=target_shape, mode=mode, mode_cval=mode_cval)
def transform(patch, flip=False, mirror=False, rotations=[]):
"""Perform data augmentation on a patch.
Args:
patch (numpy array): The patch to be processed.
flip (bool, optional): Up/down symetry.
mirror (bool, optional): left/right symetry.
rotations (int list, optional) : rotations to perform (angles in deg).
Returns:
array list: list of augmented patches
"""
transformed_patches = [patch]
for angle in rotations:
transformed_patches.append(skimage.img_as_ubyte(skimage.transform.rotate(patch, angle)))
if flip:
transformed_patches.append(np.flipud(patch))
if mirror:
transformed_patches.append(np.fliplr(patch))
return transformed_patches
# In[4]:
def load_image2(path, height=None, width=None):
# load image
img = skimage.io.imread(path)
img = img / 255.0
if height is not None and width is not None:
ny = height
nx = width
elif height is not None:
ny = height
nx = img.shape[1] * ny / img.shape[0]
elif width is not None:
nx = width
ny = img.shape[0] * nx / img.shape[1]
else:
ny = img.shape[0]
nx = img.shape[1]
return skimage.transform.resize(img, (ny, nx))
def load_image(path, size=224):
# load image
img = skimage.io.imread(path)
img = img / 255.0
assert (0 <= img).all() and (img <= 1.0).all()
# print "Original Image Shape: ", img.shape
# we crop image from center
short_edge = min(img.shape[:2])
yy = int((img.shape[0] - short_edge) / 2)
xx = int((img.shape[1] - short_edge) / 2)
crop_img = img[yy: yy + short_edge, xx: xx + short_edge]
# resize to 224, 224
resized_img = skimage.transform.resize(crop_img, (size, size))
return resized_img
# returns the top1 string
def load_image2(path, height=None, width=None):
# load image
img = skimage.io.imread(path)
img = img / 255.0
if height is not None and width is not None:
ny = height
nx = width
elif height is not None:
ny = height
nx = img.shape[1] * ny / img.shape[0]
elif width is not None:
nx = width
ny = img.shape[0] * nx / img.shape[1]
else:
ny = img.shape[0]
nx = img.shape[1]
return skimage.transform.resize(img, (ny, nx))
def rescale(img, input_height, input_width):
print("Original image shape:" + str(img.shape) + " and remember it should be in H, W, C!")
print("Model's input shape is %dx%d") % (input_height, input_width)
aspect = img.shape[1]/float(img.shape[0])
print("Orginal aspect ratio: " + str(aspect))
if(aspect>1):
# landscape orientation - wide image
res = int(aspect * input_height)
imgScaled = skimage.transform.resize(img, (input_width, res))
if(aspect<1):
# portrait orientation - tall image
res = int(input_width/aspect)
imgScaled = skimage.transform.resize(img, (res, input_height))
if(aspect == 1):
imgScaled = skimage.transform.resize(img, (input_width, input_height))
# pyplot.figure()
# pyplot.imshow(imgScaled)
# pyplot.axis('on')
# pyplot.title('Rescaled image')
print("New image shape:" + str(imgScaled.shape) + " in HWC")
return imgScaled
def build_rescale_transform_slow(downscale_factor, image_shape, target_shape):
"""
This mimics the skimage.transform.resize function.
The resulting image is centered.
"""
rows, cols = image_shape
trows, tcols = target_shape
col_scale = row_scale = downscale_factor
src_corners = np.array([[1, 1], [1, rows], [cols, rows]]) - 1
dst_corners = np.zeros(src_corners.shape, dtype=np.double)
# take into account that 0th pixel is at position (0.5, 0.5)
dst_corners[:, 0] = col_scale * (src_corners[:, 0] + 0.5) - 0.5
dst_corners[:, 1] = row_scale * (src_corners[:, 1] + 0.5) - 0.5
tform_ds = skimage.transform.AffineTransform()
tform_ds.estimate(src_corners, dst_corners)
# centering
shift_x = cols / (2.0 * downscale_factor) - tcols / 2.0
shift_y = rows / (2.0 * downscale_factor) - trows / 2.0
tform_shift_ds = skimage.transform.SimilarityTransform(translation=(shift_x, shift_y))
return tform_shift_ds + tform_ds
def perturb_multiscale(img, scale_factors, augmentation_params, target_shapes, rng=np.random):
"""
scale is a DOWNSCALING factor.
"""
tform_center, tform_uncenter = build_center_uncenter_transforms(img.shape)
tform_augment = random_perturbation_transform(rng=rng, **augmentation_params)
tform_augment = tform_uncenter + tform_augment + tform_center # shift to center, augment, shift back (for the rotation/shearing)
output = []
for scale, target_shape in zip(scale_factors, target_shapes):
if isinstance(scale, skimage.transform.ProjectiveTransform):
tform_rescale = scale
else:
tform_rescale = build_rescale_transform(scale, img.shape, target_shape) # also does centering
output.append(fast_warp(img, tform_rescale + tform_augment, output_shape=target_shape, mode='constant').astype('float32'))
return output
def perturb_multiscale_fixed(img, scale_factors, tform_augment, target_shapes):
"""
scale is a DOWNSCALING factor.
"""
tform_center, tform_uncenter = build_center_uncenter_transforms(img.shape)
tform_augment = tform_uncenter + tform_augment + tform_center # shift to center, augment, shift back (for the rotation/shearing)
output = []
for scale, target_shape in zip(scale_factors, target_shapes):
if isinstance(scale, skimage.transform.ProjectiveTransform):
tform_rescale = scale
else:
tform_rescale = build_rescale_transform(scale, img.shape, target_shape) # also does centering
output.append(fast_warp(img, tform_rescale + tform_augment, output_shape=target_shape, mode='constant').astype('float32'))
return output
def build_rescale_transform_slow(downscale_factor, image_shape, target_shape):
"""
This mimics the skimage.transform.resize function.
The resulting image is centered.
"""
rows, cols = image_shape
trows, tcols = target_shape
col_scale = row_scale = downscale_factor
src_corners = np.array([[1, 1], [1, rows], [cols, rows]]) - 1
dst_corners = np.zeros(src_corners.shape, dtype=np.double)
# take into account that 0th pixel is at position (0.5, 0.5)
dst_corners[:, 0] = col_scale * (src_corners[:, 0] + 0.5) - 0.5
dst_corners[:, 1] = row_scale * (src_corners[:, 1] + 0.5) - 0.5
tform_ds = skimage.transform.AffineTransform()
tform_ds.estimate(src_corners, dst_corners)
# centering
shift_x = cols / (2.0 * downscale_factor) - tcols / 2.0
shift_y = rows / (2.0 * downscale_factor) - trows / 2.0
tform_shift_ds = skimage.transform.SimilarityTransform(translation=(shift_x, shift_y))
return tform_shift_ds + tform_ds
def perturb_multiscale(img, scale_factors, augmentation_params, target_shapes, rng=np.random):
"""
scale is a DOWNSCALING factor.
"""
tform_center, tform_uncenter = build_center_uncenter_transforms(img.shape)
tform_augment = random_perturbation_transform(rng=rng, **augmentation_params)
tform_augment = tform_uncenter + tform_augment + tform_center # shift to center, augment, shift back (for the rotation/shearing)
output = []
for scale, target_shape in zip(scale_factors, target_shapes):
if isinstance(scale, skimage.transform.ProjectiveTransform):
tform_rescale = scale
else:
tform_rescale = build_rescale_transform(scale, img.shape, target_shape) # also does centering
output.append(fast_warp(img, tform_rescale + tform_augment, output_shape=target_shape, mode='constant').astype('float32'))
return output
def perturb_multiscale_fixed(img, scale_factors, tform_augment, target_shapes):
"""
scale is a DOWNSCALING factor.
"""
tform_center, tform_uncenter = build_center_uncenter_transforms(img.shape)
tform_augment = tform_uncenter + tform_augment + tform_center # shift to center, augment, shift back (for the rotation/shearing)
output = []
for scale, target_shape in zip(scale_factors, target_shapes):
if isinstance(scale, skimage.transform.ProjectiveTransform):
tform_rescale = scale
else:
tform_rescale = build_rescale_transform(scale, img.shape, target_shape) # also does centering
output.append(fast_warp(img, tform_rescale + tform_augment, output_shape=target_shape, mode='constant').astype('float32'))
return output
def load_image(path):
# Load image [height, width, depth]
img = skimage.io.imread(path) / 255.0
assert (0 <= img).all() and (img <= 1.0).all()
# Crop image from center
short_edge = min(img.shape[:2])
yy = int((img.shape[0] - short_edge) / 2)
xx = int((img.shape[1] - short_edge) / 2)
shape = list(img.shape)
crop_img = img[yy: yy + short_edge, xx: xx + short_edge]
resized_img = skimage.transform.resize(crop_img, (shape[0], shape[1]))
return resized_img, shape
# Return a resized numpy array of an image specified by its path
def load_image2(path, height=None, width=None):
# Load image
img = skimage.io.imread(path) / 255.0
if height is not None and width is not None:
ny = height
nx = width
elif height is not None:
ny = height
nx = img.shape[1] * ny / img.shape[0]
elif width is not None:
nx = width
ny = img.shape[0] * nx / img.shape[1]
else:
ny = img.shape[0]
nx = img.shape[1]
return skimage.transform.resize(img, (ny, nx))
# Render the generated image given a tensorflow session and a variable image (x)
def load_img(path, grayscale=False, resize=None, order=1):
# Load image
img = io.imread(path)
# Resize
# print('Desired resize: ' + str(resize))
if resize is not None:
img = skimage.transform.resize(img, resize, order=order,
preserve_range=True)
# print('Final resize: ' + str(img.shape))
# Color conversion
if len(img.shape)==2 and not grayscale:
img = gray2rgb(img)
elif len(img.shape)>2 and img.shape[2]==3 and grayscale:
img = rgb2gray(img)
# Return image
return img
def load_image(path):
# load image
nImgs = len(path)
rImg = np.zeros([nImgs,224,224,3])
for i in range(nImgs):
img = cv2.imread(path[i])
img = img / 255.0
assert (0 <= img).all() and (img <= 1.0).all()
# print "Original Image Shape: ", img.shape
# we crop image from center
short_edge = min(img.shape[:2])
yy = int((img.shape[0] - short_edge) / 2)
xx = int((img.shape[1] - short_edge) / 2)
crop_img = img[yy: yy + short_edge, xx: xx + short_edge]
# resize to 224, 224
resized_img = cv2.resize(img,(224,224),interpolation = cv2.INTER_CUBIC) #skimage.transform.resize(crop_img, (224, 224))
rImg[i] = resized_img
return rImg
# returns the top1 string
def load_image2(path, height=None, width=None):
# load image
img = skimage.io.imread(path)
img = img / 255.0
if height is not None and width is not None:
ny = height
nx = width
elif height is not None:
ny = height
nx = img.shape[1] * ny / img.shape[0]
elif width is not None:
nx = width
ny = img.shape[0] * nx / img.shape[1]
else:
ny = img.shape[0]
nx = img.shape[1]
return skimage.transform.resize(img, (ny, nx))
def load_img(path, grayscale=False, resize=None, order=1):
# Load image
img = io.imread(path)
# Resize
# print('Desired resize: ' + str(resize))
if resize is not None:
img = skimage.transform.resize(img, resize, order=order,
preserve_range=True)
# print('Final resize: ' + str(img.shape))
# Color conversion
if len(img.shape) == 2 and not grayscale:
img = gray2rgb(img)
elif len(img.shape) > 2 and img.shape[2] == 3 and grayscale:
img = rgb2gray(img)
# Return image
return img
def load_image(path):
# load image
img = skimage.io.imread(path)
img = img / 255.0
assert (0 <= img).all() and (img <= 1.0).all()
# print "Original Image Shape: ", img.shape
# we crop image from center
short_edge = min(img.shape[:2])
yy = int((img.shape[0] - short_edge) / 2)
xx = int((img.shape[1] - short_edge) / 2)
crop_img = img[yy: yy + short_edge, xx: xx + short_edge]
# resize to 224, 224
resized_img = skimage.transform.resize(crop_img, (224, 224))
if len(resized_img.shape)<3:
resized_img = skimage.color.gray2rgb(resized_img)
return resized_img
# returns the top1 string
def load_image2(path, height=None, width=None):
# load image
img = skimage.io.imread(path)
img = img / 255.0
if height is not None and width is not None:
ny = height
nx = width
elif height is not None:
ny = height
nx = img.shape[1] * ny / img.shape[0]
elif width is not None:
nx = width
ny = img.shape[0] * nx / img.shape[1]
else:
ny = img.shape[0]
nx = img.shape[1]
return skimage.transform.resize(img, (ny, nx))
def load_image(path):
# Load image [height, width, depth]
img = skimage.io.imread(path) / 255.0
assert (0 <= img).all() and (img <= 1.0).all()
# Crop image from center
short_edge = min(img.shape[:2])
yy = int((img.shape[0] - short_edge) / 2)
xx = int((img.shape[1] - short_edge) / 2)
shape = list(img.shape)
crop_img = img[yy: yy + short_edge, xx: xx + short_edge]
resized_img = skimage.transform.resize(crop_img, (shape[0], shape[1]))
return resized_img, shape
# Return a resized numpy array of an image specified by its path
def load_image2(path, height=None, width=None):
# Load image
img = skimage.io.imread(path) / 255.0
if height is not None and width is not None:
ny = height
nx = width
elif height is not None:
ny = height
nx = img.shape[1] * ny / img.shape[0]
elif width is not None:
nx = width
ny = img.shape[0] * nx / img.shape[1]
else:
ny = img.shape[0]
nx = img.shape[1]
return skimage.transform.resize(img, (ny, nx))
# Render the generated image given a tensorflow session and a variable image (x)
def load_image_array(image_file, image_size):
img = skimage.io.imread(image_file)
# GRAYSCALE
if len(img.shape) == 2:
img_new = np.ndarray( (img.shape[0], img.shape[1], 3), dtype = 'uint8')
img_new[:,:,0] = img
img_new[:,:,1] = img
img_new[:,:,2] = img
img = img_new
img_resized = skimage.transform.resize(img, (image_size, image_size))
# FLIP HORIZONTAL WIRH A PROBABILITY 0.5
if random.random() > 0.5:
img_resized = np.fliplr(img_resized)
return img_resized.astype('float32')
