cvtools.data_augs.augmentations module

class cvtools.data_augs.augmentations.Compose(transforms)

基类：object

Composes several augmentations together. :param transforms: list of transforms to compose. :type transforms: List[Transform]

Example

>>> augmentations.Compose([
>>>     transforms.CenterCrop(10),
>>>     transforms.ToTensor(),
>>> ])

class cvtools.data_augs.augmentations.ConvertColor(current='BGR', transform='HSV')

基类：object

class cvtools.data_augs.augmentations.ConvertFromInts

基类：object

class cvtools.data_augs.augmentations.Expand(mean)

基类：object

class cvtools.data_augs.augmentations.Lambda(lambd)

基类：object

Applies a lambda as a transform.

class cvtools.data_augs.augmentations.PhotometricDistort

基类：object

class cvtools.data_augs.augmentations.RandomBrightness(delta=32)

基类：object

class cvtools.data_augs.augmentations.RandomContrast(lower=0.5, upper=1.5)

基类：object

class cvtools.data_augs.augmentations.RandomHorMirror

基类：object

水平方向（flipping around the y-axis）镜像

class cvtools.data_augs.augmentations.RandomHue(delta=18.0)

基类：object

class cvtools.data_augs.augmentations.RandomLightingNoise

基类：object

class cvtools.data_augs.augmentations.RandomMirror(both=True)

基类：object

class cvtools.data_augs.augmentations.RandomRotate

基类：object

随机旋转0度、90度、180度、270度

class cvtools.data_augs.augmentations.RandomSampleCrop

基类：object

Crop :param img: the image being input during training :type img: Image :param boxes: the original bounding boxes in pt form :type boxes: Tensor :param labels: the class labels for each bbox :type labels: Tensor :param mode: the min and max jaccard overlaps :type mode: float tuple

返回:

(img, boxes, classes)

img (Image): the cropped image boxes (Tensor): the adjusted bounding boxes in pt form labels (Tensor): the class labels for each bbox

class cvtools.data_augs.augmentations.RandomSaturation(lower=0.5, upper=1.5)

基类：object

class cvtools.data_augs.augmentations.RandomVerMirror

基类：object

竖直方向（flipping around the x-axis）镜像

class cvtools.data_augs.augmentations.Resize(size=300)

基类：object

class cvtools.data_augs.augmentations.ResizeFilled(size=300)

基类：object

class cvtools.data_augs.augmentations.SSDAugmentation(size=300, mean=(104, 117, 123))

基类：object

class cvtools.data_augs.augmentations.SubtractMeans(mean)

基类：object

class cvtools.data_augs.augmentations.SwapChannels(swaps)

基类：object

Transforms a tensorized image by swapping the channels in the order

specified in the swap tuple.

参数:swaps (int triple) -- final order of channels eg: (2, 1, 0)

class cvtools.data_augs.augmentations.ToAbsoluteCoords

基类：object

class cvtools.data_augs.augmentations.ToCV2Image

基类：object

class cvtools.data_augs.augmentations.ToPercentCoords

基类：object

class cvtools.data_augs.augmentations.ToTensor

基类：object

cvtools.data_augs.augmentations.horizontal_mirror(image, boxes)

水平方向（flipping around the y-axis）镜像

参数:

• image (np.ndarray) -- numpy数组
• boxes (np.ndarray) -- numpy数组，nx4 or nx8

cvtools.data_augs.augmentations.intersect(box_a, box_b)

cvtools.data_augs.augmentations.jaccard_numpy(box_a, box_b)

Compute the jaccard overlap of two sets of boxes. The jaccard overlap is simply the intersection over union of two boxes. E.g.:

A ∩ B / A ∪ B = A ∩ B / (area(A) + area(B) - A ∩ B)

参数:

• box_a -- Multiple bounding boxes, Shape: [num_boxes,4]
• box_b -- Single bounding box, Shape: [4]

返回:

Shape: [box_a.shape[0], box_a.shape[1]]

返回类型:

jaccard overlap

cvtools.data_augs.augmentations.rotate_180(img, bboxes=None)

顺时针旋转180度

bboxes支持的格式

两点式: x1y1x2y2, 四点式：x1y1x2y2x3y3x4y4

参数:

• img (np.ndarray) -- the format of opencv image
• bboxes (list or np.ndarray) -- supported format two points: x1y1x2y2, four points: x1y1x2y2x3y3x4y4

cvtools.data_augs.augmentations.rotate_270(img, bboxes=None)

逆时针旋转90度

bboxes支持的格式

两点式: x1y1x2y2, 四点式：x1y1x2y2x3y3x4y4

参数:

• img (np.ndarray) -- the format of opencv image
• bboxes (list or np.ndarray) -- supported format two points: x1y1x2y2, four points: x1y1x2y2x3y3x4y4

cvtools.data_augs.augmentations.rotate_90(img, bboxes=None)

顺时针旋转90度

bboxes支持的格式

两点式: x1y1x2y2, 四点式：x1y1x2y2x3y3x4y4

参数:

• img (np.ndarray) -- the format of opencv image
• bboxes (list or np.ndarray) -- supported format two points: x1y1x2y2, four points: x1y1x2y2x3y3x4y4

cvtools.data_augs.augmentations.vertical_mirror(image, boxes)