如何像tf-object-detection-api那样在一个tensorboard选项卡中显示多张图像？

# 导入必要的库
import tensorflow as tf

# 假设您有多张图像，存储在一个名为"images"的列表中
images = [image1, image2, image3, ...]

# 创建一个Tensorboard摘要的会话
with tf.Session() as sess:
    # 创建一个摘要写入器
    writer = tf.summary.FileWriter('./logs', sess.graph)

    # 在摘要中写入多张图像
    image_summaries = []
    for i, image in enumerate(images):
        # 使用tf.summary.image函数创建图像摘要
        image_summary = tf.summary.image(f'image_{i}', tf.expand_dims(image, 0))
        image_summaries.append(image_summary)

    # 将摘要信息合并并写入TensorBoard
    merged_summary = tf.summary.merge(image_summaries)
    summary = sess.run(merged_summary)
    writer.add_summary(summary)

    # 关闭摘要写入器
    writer.close()