目标检测:SSD的数据增强算法

SSD的数据增强算法

代码地址

https://github.com/weiliu89/caffe/tree/ssd

论文地址

https://arxiv.org/abs/1512.02325

数据增强：

SSD数据增强有两种新方法：（1）expand ，左图（2）batch_sampler，右图

expand_param {
      prob: 0.5 //expand发生的概率
      max_expand_ratio: 4 //expand的扩大倍数
    }

expand是指对图像进行缩小，图像的其余区域补0，下图是expand的方法。个人认为这样做的目的是在数据处理阶段增加多尺度的信息。大object通过expand方法的处理可以变成小尺度的物体训练。提高ssd对尺度的泛化性。

annotated_data_param {//以下有7个batch_sampler
    batch_sampler {
      max_sample: 1
      max_trials: 1
    }
    batch_sampler {
      sampler {
        min_scale: 0.3
        max_scale: 1
        min_aspect_ratio: 0.5
        max_aspect_ratio: 2
      }
      sample_constraint {
        min_jaccard_overlap: 0.1
      }
      max_sample: 1
      max_trials: 50
    }
    batch_sampler {
      sampler {
        min_scale: 0.3
        max_scale: 1
        min_aspect_ratio: 0.5
        max_aspect_ratio: 2
      }
      sample_constraint {
        min_jaccard_overlap: 0.3
      }
      max_sample: 1
      max_trials: 50
    }
    batch_sampler {
      sampler {
        min_scale: 0.3
        max_scale: 1
        min_aspect_ratio: 0.5
        max_aspect_ratio: 2
      }
      sample_constraint {
        min_jaccard_overlap: 0.5
      }
      max_sample: 1
      max_trials: 50
    }
    batch_sampler {
      sampler {
        min_scale: 0.3
        max_scale: 1
        min_aspect_ratio: 0.5
        max_aspect_ratio: 2
      }
      sample_constraint {
        min_jaccard_overlap: 0.7
      }
      max_sample: 1
      max_trials: 50
    }
    batch_sampler {
      sampler {
        min_scale: 0.3
        max_scale: 1
        min_aspect_ratio: 0.5
        max_aspect_ratio: 2
      }
      sample_constraint {
        min_jaccard_overlap: 0.9
      }
      max_sample: 1
      max_trials: 50
    }
    batch_sampler {
      sampler {
        min_scale: 0.3
        max_scale: 1
        min_aspect_ratio: 0.5
        max_aspect_ratio: 2
      }
      sample_constraint {
        max_jaccard_overlap: 1
      }
      max_sample: 1
      max_trials: 50
    }

batch_sampler是对图像选取一个满足限制条件的区域（注意这个区域是随机抓取的）。限制条件就是抓取的patch和GT（Ground Truth）的IOU的值。

步骤是：先在区间[min_scale,max_sacle]内随机生成一个值，这个值作为patch的高Height，然后在[min_aspect_ratio,max_aspect_ratio]范围内生成ratio，从而得到patch的Width。到此为止patch的宽和高随机得到，然后在图像中进行一次patch，要求满足与GT的最小IOU是0.9，也就是IOU>=0.9。如果随机patch满足这个条件，那么张图会被resize到300300（**在SSD300300中**）送进网络训练。如下图。

    batch_sampler {
      sampler {
        min_scale: 0.3
        max_scale: 1
        min_aspect_ratio: 0.5
        max_aspect_ratio: 2
      }
      sample_constraint {
        min_jaccard_overlap: 0.9
      }
      max_sample: 1
      max_trials: 50
    }