Tensorflow(object detection API)训练模型时数据增强

修改config文件

train_config: {
  data_augmentation_options {
    random_horizontal_flip {
    }
  }
  data_augmentation_options {
    ssd_random_crop {
    }
  }
}

能选择的预处理操作

normalize_image
random_pixel_value_scale
random_image_scale
random_rgb_to_gray
random_adjust_brightness
random_adjust_contrast
random_adjust_hue
random_adjust_saturation
random_distort_color
random_jitter_boxes
random_crop_to_aspect_ratio
random_black_patches
rgb_to_gray
scale_boxes_to_pixel_coordinates
subtract_channel_mean
random_horizontal_flip
random_vertical_flip
random_rotation90
random_pad_image
random_crop_pad_image
random_resize_method
resize_image
ssd_random_crop
ssd_random_crop_pad
ssd_random_crop_fixed_aspect_ratio
ssd_random_crop_pad_fixed_aspect_ratio
参考https://github.com/tensorflow/models/blob/master/research/object_detection/samples/configs/ssd_mobilenet_v1_coco.config

你可能感兴趣的:(Tensorflow(object detection API)训练模型时数据增强)