piscis.transforms ================= .. py:module:: piscis.transforms Classes ------- .. autoapisummary:: piscis.transforms.RandomAugment Functions --------- .. autoapisummary:: piscis.transforms.batch_adjust piscis.transforms.adjust piscis.transforms.normalize piscis.transforms.standardize piscis.transforms.batch_voronoi_transform piscis.transforms.voronoi_transform Module Contents --------------- .. py:class:: RandomAugment(seed: int, output_size: Tuple[int, int], augment: bool = True) Transformer for random data augmentation. :Attributes: **rng** : np.random.Generator Random number generator used for generating random transformations. **output_size** : Optional[Tuple[int, int]] Output size. **flips0** : List[bool] List of booleans for flipping along axis 0. **flips1** : List[bool] List of booleans for flipping along axis 1. **scales** : List[float] List of image scales. **dxys** : List[Tuple[float, float]] List of image translation vectors. **thetas** : List[float] List of image rotation angles. **affines** : List[np.ndarray] List of affine transformation matrices. **intensity_shifts** : List[float] List of image intensity shift values. **intensity_scales** : List[float] List of image intensity scale factors. .. !! processed by numpydoc !! .. py:attribute:: rng .. py:attribute:: output_size .. py:attribute:: augment :value: True .. py:method:: apply(image: numpy.ndarray, coords: numpy.ndarray, filter_coords: bool = True, min_scale_factor: float = 0.75, max_scale_factor: float = 1.25, max_intensity_shift: float = 0.1, max_intensity_scale_factor: float = 5) -> Tuple[numpy.ndarray, numpy.ndarray] Apply random transformations to an image and its corresponding coordinates. :Parameters: **images** : np.ndarray Images to transform. **coords** : np.ndarray Coordinates to transform. **filter_coords** : bool, optional Whether to filter coordinates outside the transformed image. Default is True. **min_scale_factor** : float, optional Minimum scale factor. Default is 0.75. **max_scale_factor** : float, optional Maximum scale factor. Default is 1.25. **max_intensity_shift** : float, optional Maximum intensity shift. Intensity shifts are sampled from a uniform distribution with support on the interval [-`max_intensity_shift`, `max_intensity_shift`]. Default is 0.1. **max_intensity_scale_factor** : float, optional Maximum intensity scale factor. Intensity scale factors are sampled from a log-uniform distribution with support on the interval [1 / `max_intensity_scale_factor`, `max_intensity_scale_factor`]. Default is 5. :Returns: **image** : np.ndarray Transformed image. **coords** : np.ndarray Transformed coordinates. .. !! processed by numpydoc !! .. py:function:: batch_adjust(images: Sequence[numpy.ndarray], adjustment: Optional[str], **kwargs: Any) -> Sequence[numpy.ndarray] Batch adjust images. :Parameters: **images** : Sequence[np.ndarray] Images to adjust. **adjustment** : Optional[str] Adjustment type. Supported types are 'normalize' and 'standardize'. **\*\*kwargs** : Any Keyword arguments for the adjustment function. :Returns: **adjusted_images** : Sequence[np.ndarray] Adjusted images. .. !! processed by numpydoc !! .. py:function:: adjust(image: numpy.ndarray, adjustment: Optional[str], **kwargs: Any) -> numpy.ndarray Adjust an image. :Parameters: **image** : np.ndarray Image to adjust. **adjustment** : Optional[str] Adjustment type. Supported types are 'normalize' and 'standardize'. **\*\*kwargs** : Any Keyword arguments for the adjustment function. :Returns: **adjusted_image** : np.ndarray Adjusted image. :Raises: ValueError If the `adjustment` type is not supported. .. !! processed by numpydoc !! .. py:function:: normalize(image: numpy.ndarray, lower: float = 0, upper: float = 100, epsilon: float = 1e-07) -> numpy.ndarray Normalize an image to the range [0, 1] based on the specified percentiles. :Parameters: **image** : np.ndarray Image to normalize. **lower** : float, optional Lower percentile. Default is 0. **upper** : float, optional Upper percentile. Default is 100. **epsilon** : float, optional Small constant for numerical stability. Default is 1e-7. :Returns: **normalized_image** : np.ndarray Normalized image. .. !! processed by numpydoc !! .. py:function:: standardize(image: numpy.ndarray, epsilon: float = 1e-07) -> numpy.ndarray Standardize an image to zero mean and unit variance. :Parameters: **image** : np.ndarray Image to standardize. **epsilon** : float, optional Small constant for numerical stability. Default is 1e-7. :Returns: **standardized_image** : np.ndarray Standardized image. .. !! processed by numpydoc !! .. py:function:: batch_voronoi_transform(coords: Sequence[numpy.ndarray], output_size: Tuple[int, int] = (256, 256), dilation_iterations: int = 1, device: Optional[str] = None) -> Tuple[torch.Tensor, torch.Tensor] Batch Voronoi transform. :Parameters: **coords** : Sequence[np.ndarray] List of coordinates. **output_size** : Tuple[int, int], optional Output size. Default is (256, 256). **dilation_iterations** : int, optional Number of iterations to dilate ground truth labels. Default is 1. **device** : Optional[str], optional Desired device of returned tensors. Default is None. :Returns: **labels** : torch.Tensor Tensor where each pixel is a boolean for whether it contains a point in `coords`. **deltas** : torch.Tensor Tensor where each pixel is a vector to the nearest point in `coords`. .. !! processed by numpydoc !! .. py:function:: voronoi_transform(coords: numpy.ndarray, output_size: Tuple[int, int] = (256, 256), dilation_iterations: int = 1, device: Optional[str] = None, index_map: Optional[torch.Tensor] = None) -> Tuple[torch.Tensor, torch.Tensor] Transform a list of coordinates to generate ground truth binary labels and displacement vectors from each pixel to the nearest point via a Voronoi tessellation. Adapted from DeepCell Spots. :Parameters: **coords** : np.ndarray List of coordinates. **output_size** : Tuple[int, int], optional Output size. Default is (256, 256). **dilation_iterations** : int, optional Number of iterations to dilate ground truth labels. Default is 1. **device** : Optional[str], optional Desired device of returned tensors. Default is None. **index_map** : Optional[torch.Tensor], optional Precomputed index map. Will be generated internally if `index_map` is None. Default is None. :Returns: **labels** : torch.Tensor Tensor where each pixel is a boolean for whether it contains a point in `coords`. **deltas** : torch.Tensor Tensor where each pixel is a vector to the nearest point in `coords`. .. rubric:: References .. [R24833becb1da-1] Laubscher, Emily, et al. "vanvalenlab/deepcell-spots: Deep Learning Library for Spot Detection." GitHub, https://github.com/vanvalenlab/deepcell-spots. .. only:: latex [R24833becb1da-1]_ .. !! processed by numpydoc !!