[深度学习系列]-Albumentations数据增强方法（可结合pytorch调用）

最新推荐文章于 2025-03-12 20:05:52 发布

Monster_H777

最新推荐文章于 2025-03-12 20:05:52 发布

阅读量5.7k

点赞数 7

分类专栏：深度学习图像处理卷积神经网络文章标签：图像识别深度学习数据增强

本文链接：https://blog.csdn.net/Monster_H7/article/details/110820603

版权

深度学习同时被 3 个专栏收录

15 篇文章

订阅专栏

卷积神经网络

14 篇文章

订阅专栏

图像处理

5 篇文章

订阅专栏

Albumentations数据增强方法

1.前言-常用数据增强方法

2.使用-参考GIt example

3.Augmentations examples.及效果图的展示

1.MultiplicativeNoise

1.前言-常用数据增强方法
- Blur 模糊
- VerticalFlip 水平翻转
- HorizontalFlip 垂直翻转
- Flip 翻转
- Normalize 归一化
- Transpose 转置
- RandomCrop 随机裁剪
- RandomGamma 随机Gamma
- RandomRotate90 随机旋转90度
- Rotate旋转
- ShiftScaleRotate 平移缩放旋转
- CenterCrop 中心裁剪
- OpticalDistortion 光学畸变
- GridDistortion 网格失真
- ElasticTransform 弹性变换
- RandomGridShuffle 随机网格洗牌
- HueSaturationValue 色调饱和度值
- PadIfNeeded 填充
- RGBShift RGB平移
- RandomBrightness 随机亮度
- RandomContrast 随机对比度
- MotionBlur 运动模糊
- MedianBlur 中心模糊
- GaussianBlur 高斯模糊
- GaussNoise 高斯噪声
- CLAHE 对比度受限自适应直方图均衡
- InvertImg 反转图像
- ChannelShuffle 通道洗牌
- Cutout
- CoarseDropout
- ToFloat
- Crop 裁剪
- RandomScale 随机缩放
- LongestMaxSize
- SmallestMaxSize
- Resize缩放
- RandomSizedCrop 随机裁剪缩放
- RandomBrightnessContrast 随机亮度对比度
- RandomCropNearBBox
- ISONoise
- Solarize

2.使用-参考GIt example

from albumentations import (
    HorizontalFlip, IAAPerspective, ShiftScaleRotate, CLAHE, RandomRotate90,
    Transpose, ShiftScaleRotate, Blur, OpticalDistortion, GridDistortion, HueSaturationValue,
    IAAAdditiveGaussianNoise, GaussNoise, MotionBlur, MedianBlur, IAAPiecewiseAffine,
    IAASharpen, IAAEmboss, RandomBrightnessContrast, Flip, OneOf, Compose
)
import numpy as np

def strong_aug(p=0.5):
    return Compose([
        RandomRotate90(),
        Flip(),
        Transpose(),
        OneOf([
            IAAAdditiveGaussianNoise(),
            GaussNoise(),
        ], p=0.2),
        OneOf([
            MotionBlur(p=0.2),
            MedianBlur(blur_limit=3, p=0.1),
            Blur(blur_limit=3, p=0.1),
        ], p=0.2),
        ShiftScaleRotate(shift_limit=0.0625, scale_limit=0.2, rotate_limit=45, p=0.2),
        OneOf([
            OpticalDistortion(p=0.3),
            GridDistortion(p=0.1),
            IAAPiecewiseAffine(p=0.3),
        ], p=0.2),
        OneOf([
            CLAHE(clip_limit=2),
            IAASharpen(),
            IAAEmboss(),
            RandomBrightnessContrast(),
        ], p=0.3),
        HueSaturationValue(p=0.3),
    ], p=p)

image = np.ones((300, 300, 3), dtype=np.uint8)
mask = np.ones((300, 300), dtype=np.uint8)
whatever_data = "my name"
augmentation = strong_aug(p=0.9)
data = {"image": image, "mask": mask, "whatever_data": whatever_data, "additional": "hello"}
augmented = augmentation(**data)
image, mask, whatever_data, additional = augmented["image"], augmented["mask"], augmented["whatever_data"], augmented["additional"]

3.Augmentations examples.及效果图的展示

1.MultiplicativeNoise

Original image
MultiplicativeNoise(multiplier=0.5, p=1)
MultiplicativeNoise(multiplier=1.5, p=1)
MultiplicativeNoise(multiplier=[0.5, 1.5], per_channel=True, p=1)
MultiplicativeNoise(multiplier=[0.5, 1.5], elementwise=True, p=1)
MultiplicativeNoise(multiplier=[0.5, 1.5], elementwise=True, per_channel=True, p=1)