void ROIAlignForward_cpu_kernel()

in maskrcnn_benchmark/csrc/cpu/ROIAlign_cpu.cpp [114:219]

• 78 lines of code
• 10 McCabe index (conditional complexity)

void ROIAlignForward_cpu_kernel(
    const int nthreads,
    const T* bottom_data,
    const T& spatial_scale,
    const int channels,
    const int height,
    const int width,
    const int pooled_height,
    const int pooled_width,
    const int sampling_ratio,
    const T* bottom_rois,
    //int roi_cols,
    T* top_data) {
  //AT_ASSERT(roi_cols == 4 || roi_cols == 5);
  int roi_cols = 5;

  int n_rois = nthreads / channels / pooled_width / pooled_height;
  // (n, c, ph, pw) is an element in the pooled output
  // can be parallelized using omp
  // #pragma omp parallel for num_threads(32)
  for (int n = 0; n < n_rois; n++) {
    int index_n = n * channels * pooled_width * pooled_height;

    // roi could have 4 or 5 columns
    const T* offset_bottom_rois = bottom_rois + n * roi_cols;
    int roi_batch_ind = 0;
    if (roi_cols == 5) {
      roi_batch_ind = offset_bottom_rois[0];
      offset_bottom_rois++;
    }

    // Do not using rounding; this implementation detail is critical
    T roi_start_w = offset_bottom_rois[0] * spatial_scale;
    T roi_start_h = offset_bottom_rois[1] * spatial_scale;
    T roi_end_w = offset_bottom_rois[2] * spatial_scale;
    T roi_end_h = offset_bottom_rois[3] * spatial_scale;
    // T roi_start_w = round(offset_bottom_rois[0] * spatial_scale);
    // T roi_start_h = round(offset_bottom_rois[1] * spatial_scale);
    // T roi_end_w = round(offset_bottom_rois[2] * spatial_scale);
    // T roi_end_h = round(offset_bottom_rois[3] * spatial_scale);

    // Force malformed ROIs to be 1x1
    T roi_width = std::max(roi_end_w - roi_start_w, (T)1.);
    T roi_height = std::max(roi_end_h - roi_start_h, (T)1.);
    T bin_size_h = static_cast<T>(roi_height) / static_cast<T>(pooled_height);
    T bin_size_w = static_cast<T>(roi_width) / static_cast<T>(pooled_width);

    // We use roi_bin_grid to sample the grid and mimic integral
    int roi_bin_grid_h = (sampling_ratio > 0)
        ? sampling_ratio
        : ceil(roi_height / pooled_height); // e.g., = 2
    int roi_bin_grid_w =
        (sampling_ratio > 0) ? sampling_ratio : ceil(roi_width / pooled_width);

    // We do average (integral) pooling inside a bin
    const T count = roi_bin_grid_h * roi_bin_grid_w; // e.g. = 4

    // we want to precalculate indices and weights shared by all channels,
    // this is the key point of optimization
    std::vector<PreCalc<T>> pre_calc(
        roi_bin_grid_h * roi_bin_grid_w * pooled_width * pooled_height);
    pre_calc_for_bilinear_interpolate(
        height,
        width,
        pooled_height,
        pooled_width,
        roi_bin_grid_h,
        roi_bin_grid_w,
        roi_start_h,
        roi_start_w,
        bin_size_h,
        bin_size_w,
        roi_bin_grid_h,
        roi_bin_grid_w,
        pre_calc);

      for (int c = 0; c < channels; c++) {
      int index_n_c = index_n + c * pooled_width * pooled_height;
      const T* offset_bottom_data =
          bottom_data + (roi_batch_ind * channels + c) * height * width;
      int pre_calc_index = 0;

      for (int ph = 0; ph < pooled_height; ph++) {
        for (int pw = 0; pw < pooled_width; pw++) {
          int index = index_n_c + ph * pooled_width + pw;

          T output_val = 0.;
          for (int iy = 0; iy < roi_bin_grid_h; iy++) {
            for (int ix = 0; ix < roi_bin_grid_w; ix++) {
              PreCalc<T> pc = pre_calc[pre_calc_index];
              output_val += pc.w1 * offset_bottom_data[pc.pos1] +
                  pc.w2 * offset_bottom_data[pc.pos2] +
                  pc.w3 * offset_bottom_data[pc.pos3] +
                  pc.w4 * offset_bottom_data[pc.pos4];

              pre_calc_index += 1;
            }
          }
          output_val /= count;

          top_data[index] = output_val;
        } // for pw
      } // for ph
    } // for c
  } // for n
}