Python-用动漫小姐姐实现马赛克拼图

Python-动漫小姐姐像素上拼成新图片

晚上闲得无聊，就写了一个程序，在一张图上用动漫小姐姐代替每个像素。

效果如图：

代码：

import numpy as np
import cv2
import os
import pickle
import matplotlib.pyplot as plt


class Montage(object):

    def __init__(self, min_size=100):

        self.size = 28

        self.min_size = min_size

        self.data_path = './data.pkl'

        self.data_dict = self.read_data()

        self.data = self.convert_data(self.data_dict)

    def resize_image(self, image):

        image_shape = image.shape

        size_min = np.min(image_shape[:2])
        size_max = np.max(image_shape[:2])

        scale = float(self.min_size) / float(size_max)

        image = cv2.resize(image, dsize=(0, 0), fx=scale, fy=scale)

        return image

    def convert_data(self, data):

        new_data = []

        for value in data:
            new_data.append(value['value'])

        return np.array(new_data)

    def read_data(self, data_path='./image'):

        if not os.path.exists(self.data_path):

            dir_list = os.listdir(data_path)

            data = []

            for dir_path in dir_list:

                file_type = os.path.splitext(dir_path)[1]
                if file_type != '.jpeg':
                    continue

                image_path = os.path.join(data_path, dir_path)
                image = cv2.imread(image_path)

                B, G, R = cv2.split(image)
                B = np.mean(B)
                G = np.mean(G)
                R = np.mean(R)
                mean_value = (B, G, R)

                value = {'path': image_path, 'value': mean_value}

                data.append(value)

            with open(self.data_path, 'wb') as f:
                pickle.dump(data, f)

            return data
        # else

        with open(self.data_path, 'rb') as f:
            data = pickle.load(f)

        return data

    def Splice_index(self, image_path):

        image = cv2.imread(image_path)
        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)

        image = self.resize_image(image)
        h, w = image.shape[:2]

        all_index = []

        for i in range(h):

            new_line = []

            for j in range(w):

                value = image[i, j, :]

                res = self.data - value
                res = np.sum(np.abs(res), axis=-1)

                res_index = np.argmin(res)

                new_line.append(res_index)

            all_index.append(new_line)

        return all_index, h, w

    def Splice(self, image_path):

        all_index, h, w = self.Splice_index(image_path)

        row = []
        for i in range(h):

            line = []
            line_index = all_index[i]

            for j in range(w):

                index = line_index[j]
                name = self.data_dict[index]['path']
                line.append(cv2.imread(name))

            line = np.hstack(line)
            row.append(line)

        image = np.vstack(row)

        return image


if __name__ == "__main__":

    montage = Montage(min_size=100) # min_size越大，精度越高；但相应图片处理时间越长 
    # 一般设置为100处理时间为1~2s   
    image = montage.Splice('./0.jpg') # 传入图片路径

    plt.imshow(image)
    plt.show()

图片素材地址：

链接：https://pan.baidu.com/s/1fDCgbDH6qVPf7NTauMr2zA
提取码：d0fe