summaryrefslogtreecommitdiff
path: root/src/text_recognizer/datasets
diff options
context:
space:
mode:
authoraktersnurra <grydholm@kth.se>2020-12-02 23:48:52 +0100
committeraktersnurra <grydholm@kth.se>2020-12-02 23:48:52 +0100
commit5529e0fc9ca39e81fe0f08a54f257d32f0afe120 (patch)
treef2be992554e278857db7d56786dba54a76d439c7 /src/text_recognizer/datasets
parente3b039c9adb4bce42ede4cb682a3ae71e797539a (diff)
parent8e3985c9cde6666e4314973312135ec1c7a025b9 (diff)
Merge branch 'master' of github.com:aktersnurra/text-recognizer
Diffstat (limited to 'src/text_recognizer/datasets')
-rw-r--r--src/text_recognizer/datasets/transforms.py15
1 files changed, 14 insertions, 1 deletions
diff --git a/src/text_recognizer/datasets/transforms.py b/src/text_recognizer/datasets/transforms.py
index 1105f23..1ec23dc 100644
--- a/src/text_recognizer/datasets/transforms.py
+++ b/src/text_recognizer/datasets/transforms.py
@@ -4,7 +4,7 @@ from PIL import Image
import torch
from torch import Tensor
import torch.nn.functional as F
-from torchvision.transforms import Compose, ToPILImage, ToTensor
+from torchvision.transforms import Compose, RandomAffine, ToTensor
from text_recognizer.datasets.util import EmnistMapper
@@ -64,3 +64,16 @@ class AddTokens:
target = torch.cat([sos, target], dim=0)
return target
+
+
+class ApplyContrast:
+ """Sets everything below a threshold to zero, i.e. increase contrast."""
+
+ def __init__(self, low: float = 0.0, high: float = 0.25) -> None:
+ self.low = low
+ self.high = high
+
+ def __call__(self, x: Tensor) -> Tensor:
+ """Apply mask binary mask to input tensor."""
+ mask = x > np.random.RandomState().uniform(low=self.low, high=self.high)
+ return x * mask