Tensorflow加载带有图像标签的图像数据集

import tensorflow as tf
import pathlib

# 假设数据集存储在 'path_to_dataset' 目录下，目录结构如下：
# path_to_dataset/
#   ├── class1/
#   │   ├── img1.jpg
#   │   ├── img2.jpg
#   │   └── ...
#   ├── class2/
#   │   ├── img1.jpg
#   │   ├── img2.jpg
#   │   └── ...
#   └── ...

data_dir = pathlib.Path('path_to_dataset')

# 获取所有图像文件的路径和对应的标签
image_paths = list(data_dir.glob('*/*'))
image_paths = [str(path) for path in image_paths]
labels = [path.parent.name for path in image_paths]

# 创建标签到索引的映射
label_to_index = {label: index for index, label in enumerate(set(labels))}
labels = [label_to_index[label] for label in labels]

# 将数据转换为 TensorFlow Dataset
dataset = tf.data.Dataset.from_tensor_slices((image_paths, labels))

# 定义图像预处理函数
def load_and_preprocess_image(path, label):
    image = tf.io.read_file(path)
    image = tf.image.decode_jpeg(image, channels=3)
    image = tf.image.resize(image, [192, 192])
    image /= 255.0  # 归一化
    return image, label

# 应用预处理函数
dataset = dataset.map(load_and_preprocess_image)

# 设置批量大小和打乱数据
batch_size = 32
dataset = dataset.shuffle(buffer_size=len(image_paths)).batch(batch_size)

# 迭代数据集
for images, labels in dataset:
    # 在这里进行模型训练或其他操作
    pass