【ROI Pooling】Caffe版源码解析

1. LayerSetUp

template <typename Dtype>
void ROIPoolingLayer<Dtype>::LayerSetUp
      (const vector<Blob<Dtype>*>& bottom,
       const vector<Blob<Dtype>*>& top) 
{
  ROIPoolingParameter roi_pool_param = this->layer_param_.roi_pooling_param();
  //经过Pooling后的feature map的高  = 2
  pooled_height_ = roi_pool_param.pooled_h();
  //经过Pooling后的feature map的宽  = 2
  pooled_width_ = roi_pool_param.pooled_w();
  //输入图片与feature map之前的比值，这个feature map指roi pooling层的输入
  spatial_scale_实际上应该是最初输入的原图尺寸(比如64x64)/ROI pooling层的输入尺寸(8x8)，其实就是stride乘积的倒数(1/8)
  spatial_scale_ = roi_pool_param.spatial_scale();
}

2. Reshape

template <typename Dtype>
void ROIPoolingLayer<Dtype>::Reshape
     (const vector<Blob<Dtype>*>& bottom,
      const vector<Blob<Dtype>*>& top) 
{
  //输入的feature map的channel数  = 3
  channels_ = bottom[0]->channels();
  //输入的feature map的高  = 8
  height_ = bottom[0]->height();
  //输入的feature map的宽  = 8
  width_ = bottom[0]->width();
  //设置输出的形状NCHW，N=ROI的个数，C=channels_，H=pooled_height_，W=pooled_width_
  top[0]->Reshape(bottom[1]->num(), channels_, pooled_height_,
      pooled_width_);
  //max_idx_的形状与top一致
  max_idx_.Reshape(bottom[1]->num(), channels_, pooled_height_,
      pooled_width_);
}

3. Forward

template <typename Dtype>
void ROIPoolingLayer<Dtype>::Forward_cpu(
      const vector<Blob<Dtype>*>& bottom,
      const vector<Blob<Dtype>*>& top) 
{
  //输入有两部分组成，data和rois
  const Dtype* bottom_data = bottom[0]->cpu_data();   //feature map
  const Dtype* bottom_rois = bottom[1]->cpu_data();   //rois
  // Number of ROIs
  int num_rois = bottom[1]->num();
  int batch_size = bottom[0]->num();
  int top_count = top[0]->count();
  Dtype* top_data = top[0]->mutable_cpu_data();
  caffe_set(top_count, Dtype(-FLT_MAX), top_data);
  int* argmax_data = max_idx_.mutable_cpu_data();
  caffe_set(top_count, -1, argmax_data);
 
  // For each ROI R = [batch_index x1 y1 x2 y2]: max pool over R
  for (int n = 0; n < num_rois; ++n) {
    int roi_batch_ind = bottom_rois[0];   //这个ind没明白
    //把原图的roi坐标映射到feature map上面
    //注意！！！！！！！
    //这里是第一次量化   
    int roi_start_w = round(bottom_rois[1] * spatial_scale_);
    int roi_start_h = round(bottom_rois[2] * spatial_scale_);
    int roi_end_w = round(bottom_rois[3] * spatial_scale_);
    int roi_end_h = round(bottom_rois[4] * spatial_scale_);
    //计算每个roi在feature map上面的大小
    int roi_height = max(roi_end_h - roi_start_h + 1, 1);
    int roi_width = max(roi_end_w - roi_start_w + 1, 1);
    //pooling之后的feature map的一个值对应于pooling之前的feature map上的大小
    //注：由于roi的大小不一致，所以每次都需要计算一次
    //roi_height  roi_width对应原roi在feature map上的大小，除以输出尺寸(2 X 2)即分块数，
    //可以得到每一块输出对应的长宽
    //注意！！！！！！！！！
    //这里时第二次量化
    const Dtype bin_size_h = static_cast<Dtype>(roi_height)
                             / static_cast<Dtype>(pooled_height_);
    const Dtype bin_size_w = static_cast<Dtype>(roi_width)
                             / static_cast<Dtype>(pooled_width_);
    //找到对应的roi的feature map，如果input data的batch size为1
    //那么roi_batch_ind=0
    
    const Dtype* batch_data = bottom_data + bottom[0]->offset(roi_batch_ind);
    //上面这句没明白
//pooling的过程是针对每一个channel的，所以需要循环遍历
    for (int c = 0; c < channels_; ++c) {
      //计算output的每一个值，所以需要遍历一遍output，然后求出所有值
       //对每一块输出进行max pool操作
      for (int ph = 0; ph < pooled_height_; ++ph) {
        for (int pw = 0; pw < pooled_width_; ++pw) {
          // Compute pooling region for this output unit:
          //  start (included) = floor(ph * roi_height / pooled_height_)
          //  end (excluded) = ceil((ph + 1) * roi_height / pooled_height_)
          // 计算output上的一点对应于input上面区域的大小[hstart, wstart, hend, wend]
          //定位每块区域的位置
          int hstart = static_cast<int>(floor(static_cast<Dtype>(ph)
                                              * bin_size_h));
          int hend = static_cast<int>(ceil(static_cast<Dtype>(ph + 1)
                                           * bin_size_h));
          int wstart = static_cast<int>(floor(static_cast<Dtype>(pw)
                                              * bin_size_w));
          int wend = static_cast<int>(ceil(static_cast<Dtype>(pw + 1)
                                           * bin_size_w));
          //将映射后的区域平动到对应的位置[hstart, wstart, hend, wend]
          //控制在图片范围内
          hstart = min(max(hstart + roi_start_h, 0), height_);
          hend = min(max(hend + roi_start_h, 0), height_);
          wstart = min(max(wstart + roi_start_w, 0), width_);
          wend = min(max(wend + roi_start_w, 0), width_);
          //如果映射后的矩形框不符合
          bool is_empty = (hend <= hstart) || (wend <= wstart);
          //pool_index指的是此时计算的output的值对应于output的位置
          const int pool_index = ph * pooled_width_ + pw;
          //如果矩形不符合，此处output的值设为0，此处的对应于输入区域的最大值为-1
          if (is_empty) {
            top_data[pool_index] = 0;
            argmax_data[pool_index] = -1;
          }
          //遍历output的值对应于input的区域块
           // Max pool
          for (int h = hstart; h < hend; ++h) {
            for (int w = wstart; w < wend; ++w) {
             // 对应于input上的位置
              const int index = h * width_ + w;
              //计算区域块的最大值，保存在output对应的位置上
              //同时记录最大值的索引
              if (batch_data[index] > top_data[pool_index]) {
                top_data[pool_index] = batch_data[index];
                argmax_data[pool_index] = index;
              }
            }
          }
        }
      }
      // Increment all data pointers by one channel
      //通过offset指向下个位置，具体实现还要看看top的结构才能理解
      batch_data += bottom[0]->offset(0, 1);
      top_data += top[0]->offset(0, 1);
      argmax_data += max_idx_.offset(0, 1);
    }
    // Increment ROI data pointer
    bottom_rois += bottom[1]->offset(1);
  }
}

参考：https://blog.csdn.net/qq_22637925/article/details/79474863