Opencv 平均池化

将图片按照固定大小网格分割，网格内的像素值取网格内所有像素的平均值。我们将这种把图片使用均等大小网格分割，并求网格内代表值的操作称为池化（Pooling）。池化操作是卷积神经网络（Convolutional Neural Network）中重要的图像处理方式。平均池化按照下式定义：

v = 1/|R| * Sum_{i in R} v_i

请把大小为 128×128 的imori.jpg使用 8×8 的网格做平均池化。

python实现：

import cv2
import numpy as np

# Read image
img = cv2.imread("imori.jpg")

# Average Pooling
out = img.copy()

H, W, C = img.shape
G = 8
Nh = int(H / G)
Nw = int(W / G)

for y in range(Nh):
    for x in range(Nw):
        for c in range(C):
            out[G*y:G*(y+1), G*x:G*(x+1), c] = np.mean(out[G*y:G*(y+1), G*x:G*(x+1), c]).astype(np.int)

# Save result
cv2.imwrite("out.jpg", out)
cv2.imshow("result", out)
cv2.waitKey(0)
cv2.destroyAllWindows()

C++实现：

#include <opencv2/core.hpp>
#include <opencv2/highgui.hpp>
#include <iostream>
#include <math.h>

int main(int argc, const char* argv[]){
  cv::Mat img = cv::imread("imori.jpg", cv::IMREAD_COLOR);

  int width = img.rows;
  int height = img.cols;

  cv::Mat out = cv::Mat::zeros(height, width, CV_8UC3);

  int r = 8;
  double v = 0;

  for (int j = 0; j < height; j+=r){
    for (int i = 0; i < width; i+=r){
      for (int c = 0; c < 3; c++){
    v = 0;
    for (int _j = 0; _j < r; _j++){
      for (int _i = 0; _i < r; _i++){
        v += (double)img.at<cv::Vec3b>(j+_j, i+_i)[c];
      }
    }
    v /= (r * r);
    for (int _j = 0; _j < r; _j++){
      for (int _i = 0; _i < r; _i++){
        out.at<cv::Vec3b>(j+_j, i+_i)[c] = (uchar)v;
      }
    }
      }
    }
  }

  //cv::imwrite("out.jpg", out);
  cv::imshow("answer", out);
  cv::waitKey(0);
  cv::destroyAllWindows();

  return 0;
}

输入：

输出：