Secure your code as it's written. Use Snyk Code to scan source code in minutes - no build needed - and fix issues immediately.
module.global_operation.value = u"Manual"
module.minimum.value = 0.2
module.maximum.value = 0.8
module.run(workspace)
output = workspace.image_set.get_image("Thresholding")
actual = output.pixel_data
data = skimage.img_as_float(image.pixel_data)
data = skimage.exposure.rescale_intensity(data)
expected = numpy.zeros_like(data, dtype=numpy.bool)
expected[data > 0.2] = True
expected[data < 0.8] = False
numpy.testing.assert_array_equal(expected, actual)
def preprocessing2(I):
""" prepro 210x160x3 uint8 frame into 6400 (80x80) 1D float vector """
I = I[35:195] # crop
#print("x_t1",I.shape)
x_t1 = skimage.color.rgb2gray(I)
x_t1 = skimage.transform.resize(x_t1,(IMG_DIM,IMG_DIM))
x_t1 = skimage.exposure.rescale_intensity(x_t1, out_range=(0, 255))
#print("x_t1",x_t1.shape)
#print("x_t1",x_t1)
#x_t1 = x_t1.reshape(1, 1, x_t1.shape[0], x_t1.shape[1])
#s_t1 = np.append(x_t1, s_t1[:, :3, :, :], axis=1)
return x_t1 #flattens
Percentile scaling range
Returns
----
gray_rescale : numpy uint8 array
Percentile rescaled image.
"""
# Calculate intensity percentile range
pA, pB = np.percentile(gray, perc_range)
# Only rescale if limits are different
if pB == pA:
gray_rescale = gray
else:
gray_rescale = exposure.rescale_intensity(gray, in_range=(pA, pB))
return gray_rescale
if stretch_img:
img = stretch(img, 190)
if equalize:
cv2.equalizeHist(img.astype(np.uint8), img)
if contrast:
img = img.astype(np.float32)
if channels > 1:
for i_ch in range(channels):
img[..., i_ch] = exposure.rescale_intensity(img[...,i_ch].astype('float'), out_range=(0, 1))
else:
p2, p98 = np.percentile(img[img < 100], (2, 98))
img = exposure.rescale_intensity(img.astype('float'), in_range=(p2, p98), out_range=(0,1))
if standardize:
img = img.astype(np.float32)
if channels > 1:
for i_ch in range(channels):
img[...,i_ch] = preprocessing.scale(img[...,i_ch].astype('float'))
else:
img = preprocessing.scale(img.astype('float'))
img = img.astype(np.float32)
if img.shape[1] != img_cols or img.shape[0] != img_rows:
img = cv2.resize(img, (img_cols,img_rows))
img[img > 255] = 255
if stretch_img:
img = stretch(img, 190)
if equalize:
cv2.equalizeHist(img.astype(np.uint8), img)
if contrast:
img = img.astype(np.float32)
if channels > 1:
for i_ch in range(channels):
img[..., i_ch] = exposure.rescale_intensity(img[..., i_ch].astype('float'), out_range=(0, 1))
else:
img = exposure.rescale_intensity(img.astype('float'), out_range=(0, 1))
if standardize:
img = img.astype(np.float32)
if channels > 1:
for i_ch in range(channels):
img[..., i_ch] = preprocessing.scale(img[..., i_ch].astype('float'))
else:
img = preprocessing.scale(img.astype('float'))
img = img.astype(np.float32)
if img.shape[1] != img_cols or img.shape[0] != img_rows:
img = cv2.resize(img, (img_cols, img_rows))
if channels == 1:
img = np.expand_dims(img, axis=0)
# cv2.imshow("reuslt", logo)
# logo = logo[50:350, 50:350]
# extract Histogram of Oriented Gradients from the test image and
# predict the make of the car
(H, hogImage) = feature.hog(logo, orientations=8, pixels_per_cell=(8, 8),
cells_per_block=(2, 2), transform_sqrt=True, block_norm="L2", visualise=True)
pred = model.predict(H.reshape(1, -1))[0]
pred1 = model.predict_proba(H.reshape(1, -1))[0]
print(pred1)
#print(dir(pred))
# pred = model.predict(H)
# visualize the HOG image
hogImage = exposure.rescale_intensity(hogImage, out_range=(0, 255))
hogImage = hogImage.astype("uint8")
# cv2.imshow("HOG Image #{}".format(i + 1), hogImage)
# draw the prediction on the test image and display it
cv2.putText(image, pred.title(), (10, 35), cv2.FONT_HERSHEY_SIMPLEX, 1.0,
(0, 255, 0), 3)
cv2.imshow("Test Image #{}".format(i + 1), image)
cv2.waitKey(0)
arr: numpy array
N-dimensional array in rasterio band order (bands, rows, columns)
str_clip: int
The % of clip to apply to the stretch. Default = 2 (2 and 98)
Returns
----------
arr: numpy array with values stretched to the specified clip %
"""
s_min = str_clip
s_max = 100 - str_clip
arr_rescaled = np.zeros_like(arr)
for ii, band in enumerate(arr):
lower, upper = np.percentile(band, (s_min, s_max))
arr_rescaled[ii] = exposure.rescale_intensity(
band, in_range=(lower, upper)
)
return arr_rescaled.copy()
raise NameError(args.data + ' : No such env !' )
object_loc_env.action_alpha=ACTION_ALPHA
model_file =get_model_file(args)
# store the previous observations in replay memory
D =deque()
# get the first state by doing nothing and preprocess the image to 80x80x4
# do_nothing = np.zeros(ACTIONS)
# do_nothing[0] = 1
# sipdb.set_trace()
x_t, r_0, terminal = object_loc_env.localization_step(0)
# x_t = skimage.color.rgb2gray(x_t)
x_t = skimage.transform.resize(x_t,(img_rows,img_rows))
x_t = skimage.exposure.rescale_intensity(x_t,out_range=(0,255))
s_t=x_t
action_history_st =[np.zeros([ACTIONS+1]) for i in range(ACTION_HISTORY)]
action_history_st_1 =[np.zeros([ACTIONS+1]) for i in range(ACTION_HISTORY)]
# ipdb.set_trace()
# s_t = np.stack((x_t, x_t, x_t, x_t), axis=2)
#print (s_t.shape)
#In Keras, need to reshape
channels = 1 if len(s_t.shape) <=2 else s_t.shape[2]
# channels =channels*5 if args.multi_warp else channels
s_t = s_t.reshape(1, s_t.shape[0], s_t.shape[1], channels) #1*80*80*1 or 3
s_at=np.array(action_history_st)
s_at=np.reshape(s_at,(1,-1))
S_Ts=[s_t,s_at]
# ipdb.set_trace()
def preprocess(image):
image = skimage.color.rgb2gray(image)
image = skimage.transform.resize(image, (IMAGE_ROWS, IMAGE_COLS), mode = 'constant')
image = skimage.exposure.rescale_intensity(image, out_range=(0,255))
image = image.reshape(1, image.shape[0], image.shape[1], 1)
return image
def scale_rgb(layers, min_max, lidx):
layers_c = np.empty(layers.shape, dtype='float32')
# Rescale and blur.
for li in range(0, 3):
layer = layers[li]
layer = np.float32(rescale_intensity(layer,
in_range=(min_max[li][0],
min_max[li][1]),
out_range=(0, 1)))
layers_c[lidx[li]] = rescale_intensity(cv2.GaussianBlur(layer,
ksize=(3, 3),
sigmaX=3),
in_range=(0, 1),
out_range=(-1, 1))
return layers_c