Transformations
- class dataloader.transform.Compose(transforms)
Bases:
TransformA class that represents a composition of multiple transforms.
- transforms
A list of transforms to be applied sequentially.
- Type:
list
- class dataloader.transform.GammaCorrection(gamma=None)
Bases:
TransformApply gamma correction to an image.
- gamma
The gamma value to be used for gamma correction. If None, a random value between 0.5 and 1.5 will be used.
- Type:
float
- Returns:
The gamma-corrected image.
- Return type:
numpy.ndarray
- class dataloader.transform.GrayScale
Bases:
TransformTransform class to convert an image to grayscale.
- Returns:
Grayscale image.
- class dataloader.transform.Normalize(mean, std)
Bases:
TransformA transformation class to normalize an image.
- mean
The mean value(s) for normalization.
- Type:
float or tuple
- std
The standard deviation value(s) for normalization.
- Type:
float or tuple
- class dataloader.transform.RandomFlip(probability=0.5)
Bases:
TransformRandomly flips the input image horizontally or vertically with a given probability.
- probability
The probability of flipping the image. Defaults to 0.5.
- Type:
float
- Returns:
The flipped image.
- Return type:
numpy.ndarray
- class dataloader.transform.RandomRotation(probability=0.5, angle=None)
Bases:
TransformRandomly rotates an image by a specified angle or a random angle within a range.
- probability
The probability of rotating the image. Defaults to 0.5.
- Type:
float
- angle
The angle of rotation in degrees. If None, a random angle between -180 and 180 will be used.
- Type:
float
- Returns:
The rotated image.
- Return type:
numpy.ndarray
- class dataloader.transform.Resize(size)
Bases:
TransformA transformation class to resize an image to a specified size.
- size
The desired size of the image after resizing.
- Type:
tuple
- class dataloader.transform.Transform
Bases:
objectA base class for data transformations.
This class provides a template for implementing custom data transformations. Subclasses should override the __call__ method to define the specific transformation logic.