brightness_contrast_adjust optimization (#329)

* brightness_contrast_adjust optimization

* benchmark: Return brightness_contrast_adjust to standard behaviour

albumentations/augmentations/functional.py

@@ -1101,18 +1101,24 @@ def gauss_noise(image, gauss):
     return image + gauss
 
 
-def _brightness_contrast_adjust_non_uint(img, alpha=1, beta=0):
+@clipped
+def _brightness_contrast_adjust_non_uint(img, alpha=1, beta=0, beta_by_max=False):
+    dtype = img.dtype
     img = img.astype('float32')
 
     if alpha != 1:
         img *= alpha
     if beta != 0:
-        img += beta * np.mean(img)
+        if beta_by_max:
+            max_value = MAX_VALUES_BY_DTYPE[dtype]
+            img += beta * max_value
+        else:
+            img += beta * np.mean(img)
     return img
 
 
 @preserve_shape
-def _brightness_contrast_adjust_uint(img, alpha=1, beta=0):
+def _brightness_contrast_adjust_uint(img, alpha=1, beta=0, beta_by_max=False):
     dtype = np.dtype('uint8')
 
     max_value = MAX_VALUES_BY_DTYPE[dtype]
@@ -1122,19 +1128,21 @@ def _brightness_contrast_adjust_uint(img, alpha=1, beta=0):
     if alpha != 1:
         lut *= alpha
     if beta != 0:
-        lut += beta * np.mean(img)
+        if beta_by_max:
+            lut += beta * max_value
+        else:
+            lut += beta * np.mean(img)
 
     lut = np.clip(lut, 0, max_value).astype(dtype)
     img = cv2.LUT(img, lut)
     return img
 
 
-@clipped
-def brightness_contrast_adjust(img, alpha=1, beta=0):
+def brightness_contrast_adjust(img, alpha=1, beta=0, beta_by_max=False):
     if img.dtype == np.uint8:
-        return _brightness_contrast_adjust_uint(img, alpha, beta)
+        return _brightness_contrast_adjust_uint(img, alpha, beta, beta_by_max)
     else:
-        return _brightness_contrast_adjust_non_uint(img, alpha, beta)
+        return _brightness_contrast_adjust_non_uint(img, alpha, beta, beta_by_max)
 
 
 @clipped

albumentations/augmentations/transforms.py

@@ -1993,6 +1993,8 @@ class RandomBrightnessContrast(ImageOnlyTransform):
             If limit is a single float, the range will be (-limit, limit). Default: 0.2.
         contrast_limit ((float, float) or float): factor range for changing contrast.
             If limit is a single float, the range will be (-limit, limit). Default: 0.2.
+        brightness_by_max (Boolean): If True adjust contrast by image dtype maximum,
+            else adjust contrast by image mean.
         p (float): probability of applying the transform. Default: 0.5.
 
     Targets:
@@ -2002,13 +2004,18 @@ class RandomBrightnessContrast(ImageOnlyTransform):
         uint8, float32
     """
 
-    def __init__(self, brightness_limit=0.2, contrast_limit=0.2, always_apply=False, p=0.5):
+    def __init__(self, brightness_limit=0.2, contrast_limit=0.2, brightness_by_max=None, always_apply=False, p=0.5):
         super(RandomBrightnessContrast, self).__init__(always_apply, p)
         self.brightness_limit = to_tuple(brightness_limit)
         self.contrast_limit = to_tuple(contrast_limit)
+        self.brightness_by_max = brightness_by_max
+
+        if brightness_by_max is None:
+            DeprecationWarning('In the version 0.4.0 default behavior of RandomBrightnessContrast '
+                               'brightness_by_max will be changed to True.')
 
     def apply(self, img, alpha=1., beta=0., **params):
-        return F.brightness_contrast_adjust(img, alpha, beta)
+        return F.brightness_contrast_adjust(img, alpha, beta, self.brightness_by_max)
 
     def get_params(self):
         return {
@@ -2017,7 +2024,7 @@ def get_params(self):
         }
 
     def get_transform_init_args_names(self):
-        return ('brightness_limit', 'contrast_limit')
+        return ('brightness_limit', 'contrast_limit', 'brightness_by_max')
 
 
 class RandomBrightness(RandomBrightnessContrast):