from __future__ import annotations import math import random from typing import Any, Sequence, Tuple, cast from warnings import warn import cv2 import numpy as np from pydantic import AfterValidator, Field, field_validator, model_validator from typing_extensions import Annotated, Self from albumentations.augmentations.geometric import functional as fgeometric from albumentations.core.bbox_utils import union_of_bboxes from albumentations.core.pydantic import ( BorderModeType, InterpolationType, OnePlusIntRangeType, ZeroOneRangeType, check_0plus, check_01, ) from albumentations.core.transforms_interface import BaseTransformInitSchema, DualTransform from albumentations.core.types import ( NUM_MULTI_CHANNEL_DIMENSIONS, PAIR, ColorType, PercentType, PxType, ScalarType, ScaleFloatType, ScaleIntType, Targets, ) from . import functional as fcrops __all__ = [ "RandomCrop", "CenterCrop", "Crop", "CropNonEmptyMaskIfExists", "RandomSizedCrop", "RandomResizedCrop", "RandomCropNearBBox", "RandomSizedBBoxSafeCrop", "CropAndPad", "RandomCropFromBorders", "BBoxSafeRandomCrop", ] class CropSizeError(Exception): pass class CropInitSchema(BaseTransformInitSchema): height: int | None = Field(ge=1) width: int | None = Field(ge=1) class BaseCrop(DualTransform): _targets = (Targets.IMAGE, Targets.MASK, Targets.BBOXES, Targets.KEYPOINTS) def __init__(self, p: float = 1.0, always_apply: bool | None = None): super().__init__(p=p, always_apply=always_apply) def apply(self, img: np.ndarray, crop_coords: tuple[int, int, int, int], **params: Any) -> np.ndarray: x_min = crop_coords[0] y_min = crop_coords[1] x_max = crop_coords[2] y_max = crop_coords[3] return fcrops.crop(img, x_min=x_min, y_min=y_min, x_max=x_max, y_max=y_max) def apply_to_bboxes( self, bboxes: np.ndarray, crop_coords: tuple[int, int, int, int], **params: Any, ) -> np.ndarray: return fcrops.crop_bboxes_by_coords(bboxes, crop_coords, params["shape"]) def apply_to_keypoints( self, keypoints: np.ndarray, crop_coords: tuple[int, int, int, int], **params: Any, ) -> np.ndarray: return fcrops.crop_keypoints_by_coords(keypoints, crop_coords) @staticmethod def _clip_bbox(bbox: tuple[int, int, int, int], image_shape: tuple[int, int]) -> tuple[int, int, int, int]: height, width = image_shape[:2] x_min, y_min, x_max, y_max = bbox x_min = np.clip(x_min, 0, width) y_min = np.clip(y_min, 0, height) x_max = np.clip(x_max, x_min, width) y_max = np.clip(y_max, y_min, height) return x_min, y_min, x_max, y_max class RandomCrop(BaseCrop): """Crop a random part of the input. Args: height: height of the crop. width: width of the crop. p: probability of applying the transform. Default: 1. Targets: image, mask, bboxes, keypoints Image types: uint8, float32 """ class InitSchema(CropInitSchema): pass def __init__(self, height: int, width: int, p: float = 1.0, always_apply: bool | None = None): super().__init__(p=p, always_apply=always_apply) self.height = height self.width = width def get_params_dependent_on_data( self, params: dict[str, Any], data: dict[str, Any], ) -> dict[str, tuple[int, int, int, int]]: image_shape = params["shape"][:2] image_height, image_width = image_shape if self.height > image_height or self.width > image_width: raise CropSizeError( f"Crop size (height, width) exceeds image dimensions (height, width):" f" {(self.height, self.width)} vs {image_shape[:2]}", ) h_start = random.random() w_start = random.random() crop_coords = fcrops.get_crop_coords(image_shape, (self.height, self.width), h_start, w_start) return {"crop_coords": crop_coords} def get_transform_init_args_names(self) -> tuple[str, ...]: return "height", "width" class CenterCrop(BaseCrop): """Crop the central part of the input. This transform crops the center of the input image, mask, bounding boxes, and keypoints to the specified dimensions. It's useful when you want to focus on the central region of the input, discarding peripheral information. Args: height (int): The height of the crop. Must be greater than 0. width (int): The width of the crop. Must be greater than 0. p (float): Probability of applying the transform. Default: 1.0. Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: - If the specified crop size is larger than the input image in either dimension, it will raise a CropSizeError. - For bounding boxes and keypoints, only those that fall within the cropped area are kept, and their coordinates are adjusted to the new image size. Example: >>> import numpy as np >>> import albumentations as A >>> image = np.random.randint(0, 256, (100, 100, 3), dtype=np.uint8) >>> transform = A.Compose([ ... A.CenterCrop(height=80, width=80, p=1.0), ... ]) >>> transformed = transform(image=image) >>> transformed_image = transformed['image'] # 80x80 center crop of the original image """ class InitSchema(CropInitSchema): pass def __init__(self, height: int, width: int, p: float = 1.0, always_apply: bool | None = None): super().__init__(p=p, always_apply=always_apply) self.height = height self.width = width def get_transform_init_args_names(self) -> tuple[str, ...]: return "height", "width" def get_params_dependent_on_data( self, params: dict[str, Any], data: dict[str, Any], ) -> dict[str, tuple[int, int, int, int]]: image_shape = params["shape"][:2] crop_coords = fcrops.get_center_crop_coords(image_shape, (self.height, self.width)) return {"crop_coords": crop_coords} class Crop(BaseCrop): """Crop a specific region from the input image. This transform crops a rectangular region from the input image, mask, bounding boxes, and keypoints based on specified coordinates. It's useful when you want to extract a specific area of interest from your inputs. Args: x_min (int): Minimum x-coordinate of the crop region (left edge). Must be >= 0. Default: 0. y_min (int): Minimum y-coordinate of the crop region (top edge). Must be >= 0. Default: 0. x_max (int): Maximum x-coordinate of the crop region (right edge). Must be > x_min. Default: 1024. y_max (int): Maximum y-coordinate of the crop region (bottom edge). Must be > y_min. Default: 1024. always_apply (bool, optional): If set to True, the transform will be always applied. Default: None. p (float): Probability of applying the transform. Default: 1.0. Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: - The crop coordinates are applied as follows: x_min <= x < x_max and y_min <= y < y_max. - If the specified crop region extends beyond the image boundaries, it will be clipped to fit within the image. - For bounding boxes and keypoints, only those that fall within the cropped region are kept, and their coordinates are adjusted relative to the new image size. Example: >>> import numpy as np >>> import albumentations as A >>> image = np.random.randint(0, 256, (100, 100, 3), dtype=np.uint8) >>> transform = A.Compose([ ... A.Crop(x_min=10, y_min=20, x_max=80, y_max=90, p=1.0), ... ]) >>> transformed = transform(image=image) >>> transformed_image = transformed['image'] # 70x70 crop of the original image """ class InitSchema(BaseTransformInitSchema): x_min: Annotated[int, Field(ge=0)] y_min: Annotated[int, Field(ge=0)] x_max: Annotated[int, Field(gt=0)] y_max: Annotated[int, Field(gt=0)] @model_validator(mode="after") def validate_coordinates(self) -> Self: if not self.x_min < self.x_max: msg = "x_max must be greater than x_min" raise ValueError(msg) if not self.y_min < self.y_max: msg = "y_max must be greater than y_min" raise ValueError(msg) return self def __init__( self, x_min: int = 0, y_min: int = 0, x_max: int = 1024, y_max: int = 1024, always_apply: bool | None = None, p: float = 1.0, ): super().__init__(p=p, always_apply=always_apply) self.x_min = x_min self.y_min = y_min self.x_max = x_max self.y_max = y_max def get_transform_init_args_names(self) -> tuple[str, ...]: return "x_min", "y_min", "x_max", "y_max" def get_params_dependent_on_data( self, params: dict[str, Any], data: dict[str, Any], ) -> dict[str, tuple[int, int, int, int]]: return {"crop_coords": (self.x_min, self.y_min, self.x_max, self.y_max)} class CropNonEmptyMaskIfExists(BaseCrop): """Crop area with mask if mask is non-empty, else make random crop. This transform attempts to crop a region containing a mask (non-zero pixels). If the mask is empty or not provided, it falls back to a random crop. This is particularly useful for segmentation tasks where you want to focus on regions of interest defined by the mask. Args: height (int): Vertical size of crop in pixels. Must be > 0. width (int): Horizontal size of crop in pixels. Must be > 0. ignore_values (list of int, optional): Values to ignore in mask, `0` values are always ignored. For example, if background value is 5, set `ignore_values=[5]` to ignore it. Default: None. ignore_channels (list of int, optional): Channels to ignore in mask. For example, if background is the first channel, set `ignore_channels=[0]` to ignore it. Default: None. p (float): Probability of applying the transform. Default: 1.0. Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: - If a mask is provided, the transform will try to crop an area containing non-zero (or non-ignored) pixels. - If no suitable area is found in the mask or no mask is provided, it will perform a random crop. - The crop size (height, width) must not exceed the original image dimensions. - Bounding boxes and keypoints are also cropped along with the image and mask. Raises: ValueError: If the specified crop size is larger than the input image dimensions. Example: >>> import numpy as np >>> import albumentations as A >>> image = np.random.randint(0, 256, (100, 100, 3), dtype=np.uint8) >>> mask = np.zeros((100, 100), dtype=np.uint8) >>> mask[25:75, 25:75] = 1 # Create a non-empty region in the mask >>> transform = A.Compose([ ... A.CropNonEmptyMaskIfExists(height=50, width=50, p=1.0), ... ]) >>> transformed = transform(image=image, mask=mask) >>> transformed_image = transformed['image'] >>> transformed_mask = transformed['mask'] # The resulting crop will likely include part of the non-zero region in the mask """ class InitSchema(CropInitSchema): ignore_values: list[int] | None ignore_channels: list[int] | None def __init__( self, height: int, width: int, ignore_values: list[int] | None = None, ignore_channels: list[int] | None = None, always_apply: bool | None = None, p: float = 1.0, ): super().__init__(p=p, always_apply=always_apply) self.height = height self.width = width self.ignore_values = ignore_values self.ignore_channels = ignore_channels def _preprocess_mask(self, mask: np.ndarray) -> np.ndarray: mask_height, mask_width = mask.shape[:2] if self.ignore_values is not None: ignore_values_np = np.array(self.ignore_values) mask = np.where(np.isin(mask, ignore_values_np), 0, mask) if mask.ndim == NUM_MULTI_CHANNEL_DIMENSIONS and self.ignore_channels is not None: target_channels = np.array([ch for ch in range(mask.shape[-1]) if ch not in self.ignore_channels]) mask = np.take(mask, target_channels, axis=-1) if self.height > mask_height or self.width > mask_width: raise ValueError( f"Crop size ({self.height},{self.width}) is larger than image ({mask_height},{mask_width})", ) return mask def update_params(self, params: dict[str, Any], **kwargs: Any) -> dict[str, Any]: super().update_params(params, **kwargs) if "mask" in kwargs: mask = self._preprocess_mask(kwargs["mask"]) elif "masks" in kwargs and len(kwargs["masks"]): masks = kwargs["masks"] mask = self._preprocess_mask(np.copy(masks[0])) # need copy as we perform in-place mod afterwards for m in masks[1:]: mask |= self._preprocess_mask(m) else: msg = "Can not find mask for CropNonEmptyMaskIfExists" raise RuntimeError(msg) mask_height, mask_width = mask.shape[:2] if mask.any(): mask = mask.sum(axis=-1) if mask.ndim == NUM_MULTI_CHANNEL_DIMENSIONS else mask non_zero_yx = np.argwhere(mask) y, x = random.choice(non_zero_yx) x_min = x - random.randint(0, self.width - 1) y_min = y - random.randint(0, self.height - 1) x_min = np.clip(x_min, 0, mask_width - self.width) y_min = np.clip(y_min, 0, mask_height - self.height) else: x_min = random.randint(0, mask_width - self.width) y_min = random.randint(0, mask_height - self.height) x_max = x_min + self.width y_max = y_min + self.height crop_coords = x_min, y_min, x_max, y_max params["crop_coords"] = crop_coords return params def get_transform_init_args_names(self) -> tuple[str, ...]: return "height", "width", "ignore_values", "ignore_channels" class BaseRandomSizedCropInitSchema(BaseTransformInitSchema): size: tuple[int, int] @field_validator("size") @classmethod def check_size(cls, value: tuple[int, int]) -> tuple[int, int]: if any(x <= 0 for x in value): raise ValueError("All elements of 'size' must be positive integers.") return value class _BaseRandomSizedCrop(DualTransform): # Base class for RandomSizedCrop and RandomResizedCrop class InitSchema(BaseRandomSizedCropInitSchema): interpolation: InterpolationType = cv2.INTER_LINEAR def __init__( self, size: tuple[int, int], interpolation: int = cv2.INTER_LINEAR, always_apply: bool | None = None, p: float = 1.0, ): super().__init__(p, always_apply) self.size = size self.interpolation = interpolation def apply( self, img: np.ndarray, crop_coords: tuple[int, int, int, int], interpolation: int, **params: Any, ) -> np.ndarray: crop = fcrops.crop(img, *crop_coords) return fgeometric.resize(crop, self.size, interpolation) def apply_to_bboxes( self, bboxes: np.ndarray, crop_coords: tuple[int, int, int, int], **params: Any, ) -> np.ndarray: return fcrops.crop_bboxes_by_coords(bboxes, crop_coords, params["shape"]) def apply_to_keypoints( self, keypoints: np.ndarray, crop_coords: tuple[int, int, int, int], **params: Any, ) -> np.ndarray: # First, crop the keypoints cropped_keypoints = fcrops.crop_keypoints_by_coords(keypoints, crop_coords) # Calculate the dimensions of the crop crop_height = crop_coords[3] - crop_coords[1] crop_width = crop_coords[2] - crop_coords[0] # Calculate scaling factors scale_x = self.size[1] / crop_height scale_y = self.size[0] / crop_width # Scale the cropped keypoints return fgeometric.keypoints_scale(cropped_keypoints, scale_x, scale_y) def get_transform_init_args_names(self) -> tuple[str, ...]: return "size", "interpolation" class RandomSizedCrop(_BaseRandomSizedCrop): """Crop a random part of the input and rescale it to a specific size. This transform first crops a random portion of the input and then resizes it to a specified size. The size of the random crop is controlled by the 'min_max_height' parameter. Args: min_max_height (tuple[int, int]): Minimum and maximum height of the crop in pixels. size (tuple[int, int]): Target size for the output image, i.e. (height, width) after crop and resize. w2h_ratio (float): Aspect ratio (width/height) of crop. Default: 1.0 interpolation (OpenCV flag): Flag that is used to specify the interpolation algorithm. Should be one of: cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4. Default: cv2.INTER_LINEAR. p (float): Probability of applying the transform. Default: 1.0 Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: - The crop size is randomly selected for each execution within the range specified by 'min_max_height'. - The aspect ratio of the crop is determined by the 'w2h_ratio' parameter. - After cropping, the result is resized to the specified 'size'. - Bounding boxes that end up fully outside the cropped area will be removed. - Keypoints that end up outside the cropped area will be removed. - This transform differs from RandomResizedCrop in that it allows more control over the crop size through the 'min_max_height' parameter, rather than using a scale parameter. Mathematical Details: 1. A random crop height h is sampled from the range [min_max_height[0], min_max_height[1]]. 2. The crop width w is calculated as: w = h * w2h_ratio 3. A random location for the crop is selected within the input image. 4. The image is cropped to the size (h, w). 5. The crop is then resized to the specified 'size'. Example: >>> import numpy as np >>> import albumentations as A >>> image = np.random.randint(0, 256, (100, 100, 3), dtype=np.uint8) >>> transform = A.RandomSizedCrop( ... min_max_height=(50, 80), ... size=(64, 64), ... w2h_ratio=1.0, ... interpolation=cv2.INTER_LINEAR, ... p=1.0 ... ) >>> result = transform(image=image) >>> transformed_image = result['image'] # transformed_image will be a 64x64 image, resulting from a crop with height # between 50 and 80 pixels, and the same aspect ratio as specified by w2h_ratio, # taken from a random location in the original image and then resized. """ _targets = (Targets.IMAGE, Targets.MASK, Targets.BBOXES, Targets.KEYPOINTS) class InitSchema(BaseTransformInitSchema): interpolation: InterpolationType min_max_height: OnePlusIntRangeType w2h_ratio: Annotated[float, Field(gt=0)] width: int | None = Field( None, deprecated=( "Initializing with 'size' as an integer and a separate 'width' is deprecated. " "Please use a tuple (height, width) for the 'size' argument." ), ) height: int | None = Field( None, deprecated=( "Initializing with 'height' and 'width' is deprecated. " "Please use a tuple (height, width) for the 'size' argument." ), ) size: ScaleIntType | None @model_validator(mode="after") def process(self) -> Self: if isinstance(self.size, int): if isinstance(self.width, int): self.size = (self.size, self.width) else: msg = "If size is an integer, width as integer must be specified." raise TypeError(msg) if self.size is None: if self.height is None or self.width is None: message = "If 'size' is not provided, both 'height' and 'width' must be specified." raise ValueError(message) self.size = (self.height, self.width) return self def __init__( self, min_max_height: tuple[int, int], # NOTE @zetyquickly: when (width, height) are deprecated, make 'size' non optional size: ScaleIntType | None = None, width: int | None = None, height: int | None = None, *, w2h_ratio: float = 1.0, interpolation: int = cv2.INTER_LINEAR, always_apply: bool | None = None, p: float = 1.0, ): super().__init__(size=cast(Tuple[int, int], size), interpolation=interpolation, p=p, always_apply=always_apply) self.min_max_height = min_max_height self.w2h_ratio = w2h_ratio def get_params_dependent_on_data( self, params: dict[str, Any], data: dict[str, Any], ) -> dict[str, tuple[int, int, int, int]]: image_shape = params["shape"][:2] crop_height = random.randint(*self.min_max_height) crop_width = int(crop_height * self.w2h_ratio) crop_shape = (crop_height, crop_width) h_start = random.random() w_start = random.random() crop_coords = fcrops.get_crop_coords(image_shape, crop_shape, h_start, w_start) return {"crop_coords": crop_coords} def get_transform_init_args_names(self) -> tuple[str, ...]: return (*super().get_transform_init_args_names(), "min_max_height", "w2h_ratio") class RandomResizedCrop(_BaseRandomSizedCrop): """Crop a random part of the input and rescale it to a specified size. This transform first crops a random portion of the input image (or mask, bounding boxes, keypoints) and then resizes the crop to a specified size. It's particularly useful for training neural networks on images of varying sizes and aspect ratios. Args: size (tuple[int, int]): Target size for the output image, i.e. (height, width) after crop and resize. scale (tuple[float, float]): Range of the random size of the crop relative to the input size. For example, (0.08, 1.0) means the crop size will be between 8% and 100% of the input size. Default: (0.08, 1.0) ratio (tuple[float, float]): Range of aspect ratios of the random crop. For example, (0.75, 1.3333) allows crop aspect ratios from 3:4 to 4:3. Default: (0.75, 1.3333333333333333) interpolation (OpenCV flag): Flag that is used to specify the interpolation algorithm. Should be one of: cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4. Default: cv2.INTER_LINEAR p (float): Probability of applying the transform. Default: 1.0 Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: - This transform attempts to crop a random area with an aspect ratio and relative size specified by 'ratio' and 'scale' parameters. If it fails to find a suitable crop after 10 attempts, it will return a crop from the center of the image. - The crop's aspect ratio is defined as width / height. - Bounding boxes that end up fully outside the cropped area will be removed. - Keypoints that end up outside the cropped area will be removed. - After cropping, the result is resized to the specified size. Mathematical Details: 1. A target area A is sampled from the range [scale[0] * input_area, scale[1] * input_area]. 2. A target aspect ratio r is sampled from the range [ratio[0], ratio[1]]. 3. The crop width and height are computed as: w = sqrt(A * r) h = sqrt(A / r) 4. If w and h are within the input image dimensions, the crop is accepted. Otherwise, steps 1-3 are repeated (up to 10 times). 5. If no valid crop is found after 10 attempts, a centered crop is taken. 6. The crop is then resized to the specified size. Example: >>> import numpy as np >>> import albumentations as A >>> image = np.random.randint(0, 256, (100, 100, 3), dtype=np.uint8) >>> transform = A.RandomResizedCrop(size=80, scale=(0.5, 1.0), ratio=(0.75, 1.33), p=1.0) >>> result = transform(image=image) >>> transformed_image = result['image'] # transformed_image will be a 80x80 crop from a random location in the original image, # with the crop's size between 50% and 100% of the original image size, # and the crop's aspect ratio between 3:4 and 4:3. """ _targets = (Targets.IMAGE, Targets.MASK, Targets.BBOXES, Targets.KEYPOINTS) class InitSchema(BaseTransformInitSchema): scale: Annotated[tuple[float, float], AfterValidator(check_01)] ratio: Annotated[tuple[float, float], AfterValidator(check_0plus)] width: int | None = Field( None, deprecated="Initializing with 'height' and 'width' is deprecated. Use size instead.", ) height: int | None = Field( None, deprecated="Initializing with 'height' and 'width' is deprecated. Use size instead.", ) size: ScaleIntType | None interpolation: InterpolationType @model_validator(mode="after") def process(self) -> Self: if isinstance(self.size, int): if isinstance(self.width, int): self.size = (self.size, self.width) else: msg = "If size is an integer, width as integer must be specified." raise TypeError(msg) if self.size is None: if self.height is None or self.width is None: message = "If 'size' is not provided, both 'height' and 'width' must be specified." raise ValueError(message) self.size = (self.height, self.width) return self def __init__( self, # NOTE @zetyquickly: when (width, height) are deprecated, make 'size' non optional size: ScaleIntType | None = None, width: int | None = None, height: int | None = None, *, scale: tuple[float, float] = (0.08, 1.0), ratio: tuple[float, float] = (0.75, 1.3333333333333333), interpolation: int = cv2.INTER_LINEAR, always_apply: bool | None = None, p: float = 1.0, ): super().__init__(size=cast(Tuple[int, int], size), interpolation=interpolation, p=p, always_apply=always_apply) self.scale = scale self.ratio = ratio def get_params_dependent_on_data( self, params: dict[str, Any], data: dict[str, Any], ) -> dict[str, tuple[int, int, int, int]]: image_shape = params["shape"][:2] image_height, image_width = image_shape area = image_height * image_width for _ in range(10): target_area = random.uniform(*self.scale) * area log_ratio = (math.log(self.ratio[0]), math.log(self.ratio[1])) aspect_ratio = math.exp(random.uniform(*log_ratio)) width = int(round(math.sqrt(target_area * aspect_ratio))) height = int(round(math.sqrt(target_area / aspect_ratio))) if 0 < width <= image_width and 0 < height <= image_height: i = random.randint(0, image_height - height) j = random.randint(0, image_width - width) h_start = i * 1.0 / (image_height - height + 1e-10) w_start = j * 1.0 / (image_width - width + 1e-10) crop_shape = (height, width) crop_coords = fcrops.get_crop_coords(image_shape, crop_shape, h_start, w_start) return {"crop_coords": crop_coords} # Fallback to central crop in_ratio = image_width / image_height if in_ratio < min(self.ratio): width = image_width height = int(round(image_width / min(self.ratio))) elif in_ratio > max(self.ratio): height = image_height width = int(round(height * max(self.ratio))) else: # whole image width = image_width height = image_height i = (image_height - height) // 2 j = (image_width - width) // 2 h_start = i * 1.0 / (image_height - height + 1e-10) w_start = j * 1.0 / (image_width - width + 1e-10) crop_shape = (height, width) crop_coords = fcrops.get_crop_coords(image_shape, crop_shape, h_start, w_start) return {"crop_coords": crop_coords} def get_transform_init_args_names(self) -> tuple[str, ...]: return "size", "scale", "ratio", "interpolation" class RandomCropNearBBox(BaseCrop): """Crop bbox from image with random shift by x,y coordinates Args: max_part_shift (float, (float, float)): Max shift in `height` and `width` dimensions relative to `cropping_bbox` dimension. If max_part_shift is a single float, the range will be (0, max_part_shift). Default (0, 0.3). cropping_bbox_key (str): Additional target key for cropping box. Default `cropping_bbox`. cropping_box_key (str): [Deprecated] Use `cropping_bbox_key` instead. p (float): probability of applying the transform. Default: 1. Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Examples: >>> aug = Compose([RandomCropNearBBox(max_part_shift=(0.1, 0.5), cropping_bbox_key='test_bbox')], >>> bbox_params=BboxParams("pascal_voc")) >>> result = aug(image=image, bboxes=bboxes, test_bbox=[0, 5, 10, 20]) """ _targets = (Targets.IMAGE, Targets.MASK, Targets.BBOXES, Targets.KEYPOINTS) class InitSchema(BaseTransformInitSchema): max_part_shift: ZeroOneRangeType cropping_bbox_key: str def __init__( self, max_part_shift: ScaleFloatType = (0, 0.3), cropping_bbox_key: str = "cropping_bbox", cropping_box_key: str | None = None, # Deprecated always_apply: bool | None = None, p: float = 1.0, ): super().__init__(p=p, always_apply=always_apply) # Check for deprecated parameter and issue warning if cropping_box_key is not None: warn( "The parameter 'cropping_box_key' is deprecated and will be removed in future versions. " "Use 'cropping_bbox_key' instead.", DeprecationWarning, stacklevel=2, ) # Ensure the new parameter is used even if the old one is passed cropping_bbox_key = cropping_box_key self.max_part_shift = cast(Tuple[float, float], max_part_shift) self.cropping_bbox_key = cropping_bbox_key def get_params_dependent_on_data( self, params: dict[str, Any], data: dict[str, Any], ) -> dict[str, tuple[float, ...]]: bbox = data[self.cropping_bbox_key] image_shape = params["shape"][:2] bbox = self._clip_bbox(bbox, image_shape) h_max_shift = round((bbox[3] - bbox[1]) * self.max_part_shift[0]) w_max_shift = round((bbox[2] - bbox[0]) * self.max_part_shift[1]) x_min = bbox[0] - random.randint(-w_max_shift, w_max_shift) x_max = bbox[2] + random.randint(-w_max_shift, w_max_shift) y_min = bbox[1] - random.randint(-h_max_shift, h_max_shift) y_max = bbox[3] + random.randint(-h_max_shift, h_max_shift) crop_coords = self._clip_bbox((x_min, y_min, x_max, y_max), image_shape) if crop_coords[0] == crop_coords[2] or crop_coords[1] == crop_coords[3]: crop_shape = (bbox[3] - bbox[1], bbox[2] - bbox[0]) crop_coords = fcrops.get_center_crop_coords(image_shape, crop_shape) return {"crop_coords": crop_coords} @property def targets_as_params(self) -> list[str]: return [self.cropping_bbox_key] def get_transform_init_args_names(self) -> tuple[str, ...]: return "max_part_shift", "cropping_bbox_key" class BBoxSafeRandomCrop(BaseCrop): """Crop a random part of the input without loss of bounding boxes. This transform performs a random crop of the input image while ensuring that all bounding boxes remain within the cropped area. It's particularly useful for object detection tasks where preserving all objects in the image is crucial. Args: erosion_rate (float): A value between 0.0 and 1.0 that determines the minimum allowable size of the crop as a fraction of the original image size. For example, an erosion_rate of 0.2 means the crop will be at least 80% of the original image height. Default: 0.0 (no minimum size). p (float): Probability of applying the transform. Default: 1.0. Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: This transform ensures that all bounding boxes in the original image are fully contained within the cropped area. If it's not possible to find such a crop (e.g., when bounding boxes are too spread out), it will default to cropping the entire image. Example: >>> import numpy as np >>> import albumentations as A >>> image = np.ones((300, 300, 3), dtype=np.uint8) >>> bboxes = [(10, 10, 50, 50), (100, 100, 150, 150)] >>> transform = A.Compose([ ... A.BBoxSafeRandomCrop(erosion_rate=0.2, p=1.0), ... ], bbox_params=A.BboxParams(format='pascal_voc', label_fields=['labels'])) >>> transformed = transform(image=image, bboxes=bboxes, labels=['cat', 'dog']) >>> transformed_image = transformed['image'] >>> transformed_bboxes = transformed['bboxes'] """ _targets = (Targets.IMAGE, Targets.MASK, Targets.BBOXES, Targets.KEYPOINTS) class InitSchema(BaseTransformInitSchema): erosion_rate: float = Field( ge=0.0, le=1.0, ) def __init__(self, erosion_rate: float = 0.0, p: float = 1.0, always_apply: bool | None = None): super().__init__(p=p, always_apply=always_apply) self.erosion_rate = erosion_rate def _get_coords_no_bbox(self, image_shape: tuple[int, int]) -> tuple[int, int, int, int]: image_height, image_width = image_shape erosive_h = int(image_height * (1.0 - self.erosion_rate)) crop_height = image_height if erosive_h >= image_height else random.randint(erosive_h, image_height) crop_width = int(crop_height * image_width / image_height) h_start = random.random() w_start = random.random() crop_shape = (crop_height, crop_width) return fcrops.get_crop_coords(image_shape, crop_shape, h_start, w_start) def get_params_dependent_on_data( self, params: dict[str, Any], data: dict[str, Any], ) -> dict[str, tuple[int, int, int, int]]: image_shape = params["shape"][:2] if len(data["bboxes"]) == 0: # less likely, this class is for use with bboxes. crop_coords = self._get_coords_no_bbox(image_shape) return {"crop_coords": crop_coords} bbox_union = union_of_bboxes(bboxes=data["bboxes"], erosion_rate=self.erosion_rate) if bbox_union is None: crop_coords = self._get_coords_no_bbox(image_shape) return {"crop_coords": crop_coords} x_min, y_min, x_max, y_max = bbox_union x_min = np.clip(x_min, 0, 1) y_min = np.clip(y_min, 0, 1) x_max = np.clip(x_max, x_min, 1) y_max = np.clip(y_max, y_min, 1) image_height, image_width = image_shape crop_x_min = int(x_min * random.random() * image_width) crop_y_min = int(y_min * random.random() * image_height) bbox_xmax = x_max + (1 - x_max) * random.random() bbox_ymax = y_max + (1 - y_max) * random.random() crop_x_max = int(bbox_xmax * image_width) crop_y_max = int(bbox_ymax * image_height) return {"crop_coords": (crop_x_min, crop_y_min, crop_x_max, crop_y_max)} @property def targets_as_params(self) -> list[str]: return ["bboxes"] def get_transform_init_args_names(self) -> tuple[str, ...]: return ("erosion_rate",) class RandomSizedBBoxSafeCrop(BBoxSafeRandomCrop): """Crop a random part of the input and rescale it to a specific size without loss of bounding boxes. This transform first attempts to crop a random portion of the input image while ensuring that all bounding boxes remain within the cropped area. It then resizes the crop to the specified size. This is particularly useful for object detection tasks where preserving all objects in the image is crucial while also standardizing the image size. Args: height (int): Height of the output image after resizing. width (int): Width of the output image after resizing. erosion_rate (float): A value between 0.0 and 1.0 that determines the minimum allowable size of the crop as a fraction of the original image size. For example, an erosion_rate of 0.2 means the crop will be at least 80% of the original image height and width. Default: 0.0 (no minimum size). interpolation (OpenCV flag): Flag that is used to specify the interpolation algorithm. Should be one of: cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.AREA, cv2.INTER_LANCZOS4. Default: cv2.INTER_LINEAR. p (float): Probability of applying the transform. Default: 1.0. Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: - This transform ensures that all bounding boxes in the original image are fully contained within the cropped area. If it's not possible to find such a crop (e.g., when bounding boxes are too spread out), it will default to cropping the entire image. - After cropping, the result is resized to the specified (height, width) size. - Bounding box coordinates are adjusted to match the new image size. - Keypoints are moved along with the crop and scaled to the new image size. - If there are no bounding boxes in the image, it will fall back to a random crop. Mathematical Details: 1. A crop region is selected that includes all bounding boxes. 2. The crop size is determined by the erosion_rate: min_crop_size = (1 - erosion_rate) * original_size 3. If the selected crop is smaller than min_crop_size, it's expanded to meet this requirement. 4. The crop is then resized to the specified (height, width) size. 5. Bounding box coordinates are transformed to match the new image size: new_coord = (old_coord - crop_start) * (new_size / crop_size) Example: >>> import numpy as np >>> import albumentations as A >>> image = np.random.randint(0, 256, (300, 300, 3), dtype=np.uint8) >>> bboxes = [(10, 10, 50, 50), (100, 100, 150, 150)] >>> transform = A.Compose([ ... A.RandomSizedBBoxSafeCrop(height=224, width=224, erosion_rate=0.2, p=1.0), ... ], bbox_params=A.BboxParams(format='pascal_voc', label_fields=['labels'])) >>> transformed = transform(image=image, bboxes=bboxes, labels=['cat', 'dog']) >>> transformed_image = transformed['image'] >>> transformed_bboxes = transformed['bboxes'] # transformed_image will be a 224x224 image containing all original bounding boxes, # with their coordinates adjusted to the new image size. """ _targets = (Targets.IMAGE, Targets.MASK, Targets.BBOXES, Targets.KEYPOINTS) class InitSchema(CropInitSchema): erosion_rate: float = Field( ge=0.0, le=1.0, ) interpolation: InterpolationType def __init__( self, height: int, width: int, erosion_rate: float = 0.0, interpolation: int = cv2.INTER_LINEAR, always_apply: bool | None = None, p: float = 1.0, ): super().__init__(erosion_rate=erosion_rate, p=p, always_apply=always_apply) self.height = height self.width = width self.interpolation = interpolation def apply( self, img: np.ndarray, crop_coords: tuple[int, int, int, int], **params: Any, ) -> np.ndarray: crop = fcrops.crop(img, *crop_coords) return fgeometric.resize(crop, (self.height, self.width), self.interpolation) def apply_to_keypoint( self, keypoints: np.ndarray, crop_coords: tuple[int, int, int, int], **params: Any, ) -> np.ndarray: keypoints = fcrops.crop_keypoints_by_coords(keypoints, crop_coords) crop_height = crop_coords[3] - crop_coords[1] crop_width = crop_coords[2] - crop_coords[0] scale_y = self.height / crop_height scale_x = self.width / crop_width return fgeometric.keypoints_scale(keypoints, scale_x=scale_x, scale_y=scale_y) def get_transform_init_args_names(self) -> tuple[str, ...]: return (*super().get_transform_init_args_names(), "height", "width", "interpolation") class CropAndPad(DualTransform): """Crop and pad images by pixel amounts or fractions of image sizes. This transform allows for simultaneous cropping and padding of images. Cropping removes pixels from the sides (i.e., extracts a subimage), while padding adds pixels to the sides (e.g., black pixels). The amount of cropping/padding can be specified either in absolute pixels or as a fraction of the image size. Args: px (int, tuple of int, tuple of tuples of int, or None): The number of pixels to crop (negative values) or pad (positive values) on each side of the image. Either this or the parameter `percent` may be set, not both at the same time. - If int: crop/pad all sides by this value. - If tuple of 2 ints: crop/pad by (top/bottom, left/right). - If tuple of 4 ints: crop/pad by (top, right, bottom, left). - Each int can also be a tuple of 2 ints for a range, or a list of ints for discrete choices. Default: None. percent (float, tuple of float, tuple of tuples of float, or None): The fraction of the image size to crop (negative values) or pad (positive values) on each side. Either this or the parameter `px` may be set, not both at the same time. - If float: crop/pad all sides by this fraction. - If tuple of 2 floats: crop/pad by (top/bottom, left/right) fractions. - If tuple of 4 floats: crop/pad by (top, right, bottom, left) fractions. - Each float can also be a tuple of 2 floats for a range, or a list of floats for discrete choices. Default: None. pad_mode (int): OpenCV border mode used for padding. Default: cv2.BORDER_CONSTANT. pad_cval (number, tuple of number, or list of number): The constant value to use for padding if pad_mode is cv2.BORDER_CONSTANT. - If number: use this value for all channels. - If tuple of 2 numbers: use uniform random value between these numbers. - If list of numbers: use random choice from this list. Default: 0. pad_cval_mask (number, tuple of number, or list of number): Same as pad_cval but used for mask padding. Default: 0. keep_size (bool): If True, the output image will be resized to the input image size after cropping/padding. Default: True. sample_independently (bool): If True and ranges are used for px/percent, sample a value for each side independently. If False, sample one value and use it for all sides. Default: True. interpolation (int): OpenCV interpolation flag used for resizing if keep_size is True. Default: cv2.INTER_LINEAR. p (float): Probability of applying the transform. Default: 1.0. Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: - This transform will never crop images below a height or width of 1. - When using pixel values (px), the image will be cropped/padded by exactly that many pixels. - When using percentages (percent), the amount of crop/pad will be calculated based on the image size. - Bounding boxes that end up fully outside the image after cropping will be removed. - Keypoints that end up outside the image after cropping will be removed. Example: >>> import albumentations as A >>> transform = A.Compose([ ... A.CropAndPad(px=(-10, 20, 30, -40), pad_mode=cv2.BORDER_REFLECT, p=1.0), ... ]) >>> transformed = transform(image=image, mask=mask, bboxes=bboxes, keypoints=keypoints) >>> transformed_image = transformed['image'] >>> transformed_mask = transformed['mask'] >>> transformed_bboxes = transformed['bboxes'] >>> transformed_keypoints = transformed['keypoints'] """ _targets = (Targets.IMAGE, Targets.MASK, Targets.BBOXES, Targets.KEYPOINTS) class InitSchema(BaseTransformInitSchema): px: PxType | None percent: PercentType | None pad_mode: BorderModeType pad_cval: ScalarType | tuple[ScalarType, ScalarType] | list[ScalarType] pad_cval_mask: ScalarType | tuple[ScalarType, ScalarType] | list[ScalarType] keep_size: bool sample_independently: bool interpolation: InterpolationType @model_validator(mode="after") def check_px_percent(self) -> Self: if self.px is None and self.percent is None: msg = "Both px and percent parameters cannot be None simultaneously." raise ValueError(msg) if self.px is not None and self.percent is not None: msg = "Only px or percent may be set!" raise ValueError(msg) return self def __init__( self, px: int | list[int] | None = None, percent: float | list[float] | None = None, pad_mode: int = cv2.BORDER_CONSTANT, pad_cval: ScalarType | tuple[ScalarType, ScalarType] | list[ScalarType] = 0, pad_cval_mask: ScalarType | tuple[ScalarType, ScalarType] | list[ScalarType] = 0, keep_size: bool = True, sample_independently: bool = True, interpolation: int = cv2.INTER_LINEAR, always_apply: bool | None = None, p: float = 1.0, ): super().__init__(p=p, always_apply=always_apply) self.px = px self.percent = percent self.pad_mode = pad_mode self.pad_cval = pad_cval self.pad_cval_mask = pad_cval_mask self.keep_size = keep_size self.sample_independently = sample_independently self.interpolation = interpolation def apply( self, img: np.ndarray, crop_params: Sequence[int], pad_params: Sequence[int], pad_value: ColorType, interpolation: int, **params: Any, ) -> np.ndarray: return fcrops.crop_and_pad( img, crop_params, pad_params, pad_value, params["shape"][:2], interpolation, self.pad_mode, self.keep_size, ) def apply_to_mask( self, mask: np.ndarray, crop_params: Sequence[int], pad_params: Sequence[int], pad_value_mask: float, interpolation: int, **params: Any, ) -> np.ndarray: return fcrops.crop_and_pad( mask, crop_params, pad_params, pad_value_mask, params["shape"][:2], interpolation, self.pad_mode, self.keep_size, ) def apply_to_bboxes( self, bboxes: np.ndarray, crop_params: tuple[int, int, int, int], pad_params: tuple[int, int, int, int], result_shape: tuple[int, int], **params: Any, ) -> np.ndarray: return fcrops.crop_and_pad_bboxes(bboxes, crop_params, pad_params, params["shape"][:2], result_shape) def apply_to_keypoints( self, keypoints: np.ndarray, crop_params: tuple[int, int, int, int], pad_params: tuple[int, int, int, int], result_shape: tuple[int, int], **params: Any, ) -> np.ndarray: return fcrops.crop_and_pad_keypoints( keypoints, crop_params, pad_params, params["shape"][:2], result_shape, self.keep_size, ) @staticmethod def __prevent_zero(val1: int, val2: int, max_val: int) -> tuple[int, int]: regain = abs(max_val) + 1 regain1 = regain // 2 regain2 = regain // 2 if regain1 + regain2 < regain: regain1 += 1 if regain1 > val1: diff = regain1 - val1 regain1 = val1 regain2 += diff elif regain2 > val2: diff = regain2 - val2 regain2 = val2 regain1 += diff return val1 - regain1, val2 - regain2 @staticmethod def _prevent_zero(crop_params: list[int], height: int, width: int) -> list[int]: top, right, bottom, left = crop_params remaining_height = height - (top + bottom) remaining_width = width - (left + right) if remaining_height < 1: top, bottom = CropAndPad.__prevent_zero(top, bottom, height) if remaining_width < 1: left, right = CropAndPad.__prevent_zero(left, right, width) return [max(top, 0), max(right, 0), max(bottom, 0), max(left, 0)] def get_params_dependent_on_data(self, params: dict[str, Any], data: dict[str, Any]) -> dict[str, Any]: height, width = params["shape"][:2] if self.px is not None: new_params = self._get_px_params() else: percent_params = self._get_percent_params() new_params = [ int(percent_params[0] * height), int(percent_params[1] * width), int(percent_params[2] * height), int(percent_params[3] * width), ] pad_params = [max(i, 0) for i in new_params] crop_params = self._prevent_zero([-min(i, 0) for i in new_params], height, width) top, right, bottom, left = crop_params crop_params = [left, top, width - right, height - bottom] result_rows = crop_params[3] - crop_params[1] result_cols = crop_params[2] - crop_params[0] if result_cols == width and result_rows == height: crop_params = [] top, right, bottom, left = pad_params pad_params = [top, bottom, left, right] if any(pad_params): result_rows += top + bottom result_cols += left + right else: pad_params = [] return { "crop_params": crop_params or None, "pad_params": pad_params or None, "pad_value": None if pad_params is None else self._get_pad_value(self.pad_cval), "pad_value_mask": None if pad_params is None else self._get_pad_value(self.pad_cval_mask), "result_shape": (result_rows, result_cols), } def _get_px_params(self) -> list[int]: if self.px is None: msg = "px is not set" raise ValueError(msg) if isinstance(self.px, int): params = [self.px] * 4 elif len(self.px) == PAIR: if self.sample_independently: params = [random.randrange(*self.px) for _ in range(4)] else: px = random.randrange(*self.px) params = [px] * 4 elif isinstance(self.px[0], int): params = self.px elif len(self.px[0]) == PAIR: params = [random.randrange(*i) for i in self.px] else: params = [random.choice(i) for i in self.px] return params def _get_percent_params(self) -> list[float]: if self.percent is None: msg = "percent is not set" raise ValueError(msg) if isinstance(self.percent, float): params = [self.percent] * 4 elif len(self.percent) == PAIR: if self.sample_independently: params = [random.uniform(*self.percent) for _ in range(4)] else: px = random.uniform(*self.percent) params = [px] * 4 elif isinstance(self.percent[0], (int, float)): params = self.percent elif len(self.percent[0]) == PAIR: params = [random.uniform(*i) for i in self.percent] else: params = [random.choice(i) for i in self.percent] return params # params = [top, right, bottom, left] @staticmethod def _get_pad_value( pad_value: ScalarType | tuple[ScalarType, ScalarType] | list[ScalarType], ) -> ScalarType: if isinstance(pad_value, (int, float)): return pad_value if len(pad_value) == PAIR: a, b = pad_value if isinstance(a, int) and isinstance(b, int): return random.randint(a, b) return random.uniform(a, b) return random.choice(pad_value) def get_transform_init_args_names(self) -> tuple[str, ...]: return ( "px", "percent", "pad_mode", "pad_cval", "pad_cval_mask", "keep_size", "sample_independently", "interpolation", ) class RandomCropFromBorders(BaseCrop): """Randomly crops the input from its borders without resizing. This transform randomly crops parts of the input (image, mask, bounding boxes, or keypoints) from each of its borders. The amount of cropping is specified as a fraction of the input's dimensions for each side independently. Args: crop_left (float): The maximum fraction of width to crop from the left side. Must be in the range [0.0, 1.0]. Default: 0.1 crop_right (float): The maximum fraction of width to crop from the right side. Must be in the range [0.0, 1.0]. Default: 0.1 crop_top (float): The maximum fraction of height to crop from the top. Must be in the range [0.0, 1.0]. Default: 0.1 crop_bottom (float): The maximum fraction of height to crop from the bottom. Must be in the range [0.0, 1.0]. Default: 0.1 p (float): Probability of applying the transform. Default: 1.0 Targets: image, mask, bboxes, keypoints Image types: uint8, float32 Note: - The actual amount of cropping for each side is randomly chosen between 0 and the specified maximum for each application of the transform. - The sum of crop_left and crop_right must not exceed 1.0, and the sum of crop_top and crop_bottom must not exceed 1.0. Otherwise, a ValueError will be raised. - This transform does not resize the input after cropping, so the output dimensions will be smaller than the input dimensions. - Bounding boxes that end up fully outside the cropped area will be removed. - Keypoints that end up outside the cropped area will be removed. Example: >>> import numpy as np >>> import albumentations as A >>> image = np.random.randint(0, 256, (100, 100, 3), dtype=np.uint8) >>> transform = A.RandomCropFromBorders( ... crop_left=0.1, crop_right=0.2, crop_top=0.2, crop_bottom=0.1, p=1.0 ... ) >>> result = transform(image=image) >>> transformed_image = result['image'] # The resulting image will have random crops from each border, with the maximum # possible crops being 10% from the left, 20% from the right, 20% from the top, # and 10% from the bottom. The image size will be reduced accordingly. """ _targets = (Targets.IMAGE, Targets.MASK, Targets.BBOXES, Targets.KEYPOINTS) class InitSchema(BaseTransformInitSchema): crop_left: float = Field( ge=0.0, le=1.0, ) crop_right: float = Field( ge=0.0, le=1.0, ) crop_top: float = Field( ge=0.0, le=1.0, ) crop_bottom: float = Field( ge=0.0, le=1.0, ) @model_validator(mode="after") def validate_crop_values(self) -> Self: if self.crop_left + self.crop_right > 1.0: msg = "The sum of crop_left and crop_right must be <= 1." raise ValueError(msg) if self.crop_top + self.crop_bottom > 1.0: msg = "The sum of crop_top and crop_bottom must be <= 1." raise ValueError(msg) return self def __init__( self, crop_left: float = 0.1, crop_right: float = 0.1, crop_top: float = 0.1, crop_bottom: float = 0.1, always_apply: bool | None = None, p: float = 1.0, ): super().__init__(p=p, always_apply=always_apply) self.crop_left = crop_left self.crop_right = crop_right self.crop_top = crop_top self.crop_bottom = crop_bottom def get_params_dependent_on_data( self, params: dict[str, Any], data: dict[str, Any], ) -> dict[str, tuple[int, int, int, int]]: height, width = params["shape"][:2] x_min = random.randint(0, int(self.crop_left * width)) x_max = random.randint(max(x_min + 1, int((1 - self.crop_right) * width)), width) y_min = random.randint(0, int(self.crop_top * height)) y_max = random.randint(max(y_min + 1, int((1 - self.crop_bottom) * height)), height) crop_coords = x_min, y_min, x_max, y_max return {"crop_coords": crop_coords} def get_transform_init_args_names(self) -> tuple[str, ...]: return "crop_left", "crop_right", "crop_top", "crop_bottom"