9 examples of 'add gaussian noise to image python' in Python

176
def __call__(self, image):
177
    if isinstance(self.sigma, collections.Sequence):
178
        sigma = random_num_generator(
179
            self.sigma, random_state=self.random_state)
180
    else:
181
        sigma = self.sigma
182
    if isinstance(self.mean, collections.Sequence, random_state=self.random_state):
183
        mean = random_num_generator(self.mean)
184
    else:
185
        mean = self.mean
186
    row, col, ch = image.shape
187
    gauss = self.random_state.normal(mean, sigma, (row, col, ch))
188
    gauss = gauss.reshape(row, col, ch)
189
    image += gauss
190
    return image

26
def addGaussianNoise(src):
27
    row,col,ch= src.shape
28
    mean = 0
29
    var = 0.1
30
    sigma = 15
31
    gauss = np.random.normal(mean,sigma,(row,col,ch))
32
    gauss = gauss.reshape(row,col,ch)
33
    noisy = src + gauss
34
35
    return noisy

305
def noiseImage(n1,n2):
306
  r = Random(3)
307
  x = sub(randfloat(r,n1,n2),0.5)
308
  rgf = RecursiveGaussianFilter(2.0)
309
  for x2 in x:
310
    rgf.apply1(x2,x2)
311
  return x

176
def AddGauss(img, level):
177
    return cv2.blur(img, (level * 2 + 1, level * 2 + 1));

29
def compute_scaled_noise(self, noise, background_noise):
30
    """Compute a scaled noise map from the baseline noise map. This scales each galaxy component individually \
31
    using their galaxy contribution map and sums their scaled noise maps with the baseline and background noise maps.
32
33
    Parameters
34
    -----------
35
    noise : ndarray
36
        The noise before scaling (electrons per second)..
37
    background_noise : ndarray
38
        The background noise values (electrons per second)..
39
    """
40
    return noise + (self.background_noise_scale * background_noise)

138
def _add_noise(self, img):
139
    """Adds Gaussian or Poisson noise to image."""
140
141
    w, h = img.size
142
    c = len(img.getbands())
143
144
    # Poisson distribution
145
    if self.noise_type == 'poisson':
146
        noise = np.random.poisson(self.noise_param, (h, w, c))
147
148
    # Normal distribution (default)
149
    else:
150
        std = np.random.uniform(0, self.noise_param)
151
        noise = np.random.normal(0, std, (h, w, c))
152
153
    # Add noise and clip
154
    noise_img = np.array(img) + noise
155
    noise_img = np.clip(noise_img, 0, 255).astype(np.uint8)
156
157
    return Image.fromarray(noise_img)

87
def rand_noise(img):
88
  img_noisy = img + np.random.normal(scale=noise, size=img.shape)
89
  return img_noisy

85
def remove_background_gauss(img, min_sigma=3, max_sigma=30, threshold=1):
86
    """Remove background from an image using a difference of gaussian approach
87
88
    img: ndarray
89
        Image array
90
    min_sigma: float, optional
91
        The minimum standard deviation for the gaussian filter
92
    max_sigma: float, optional
93
        The maximum standard deviation for the gaussian filter
94
    threshold: float, optional
95
        Remove any remaining features below this threshold
96
97
    Returns img: ndarray
98
        Image array with background removed
99
    """
100
    img_float = img.astype(float)
101
    img_corr = np.maximum(ndimage.gaussian_filter(img_float, min_sigma) - ndimage.gaussian_filter(img_float, max_sigma) - threshold, 0)
102
    return img_corr.astype(int)

86
def filter(image):
87
    image_ = clean_image.convert_to_greyscale(image)
88
    image_ = clean_image.clean_noise(image_, 3)
89
    return image_