如何在keras中预处理后将像素矩阵显示为图像？

Question

在 keras 中拟合神经网络之前，我正在处理我的图像。我想看看图像在预处理之后但在训练模型之前是什么样子（这样我可以确保我正确地进行了预处理）。下面的命令生成一个对象并将其存储在 train_image_array_gen 中。但是，当我尝试访问每个图像时，它们被存储为多维像素矩阵。如何直观地显示这些像素矩阵？

train_data_gen <- image_data_generator(
  rescale = 1/255
)

#train_image_files_path is just a directory of training images
train_image_array_gen <- flow_images_from_directory(train_image_files_path, 
                                                    train_data_gen, 
                                                    target_size = c(100,100), 
                                                    class_mode = "categorical", 
                                                    classes = c('0', '1', '2'), 
                                                    seed = 42, color_mode="grayscale")

print(train_image_array_gen[1])
#outputs array of pixels like so:

[[1]]
, , 1, 1

           [,1]      [,2]      [,3]      [,4]      [,5]      [,6]      [,7]      [,8]      [,9]     [,10]
 [1,] 0.2156863 0.2745098 0.3176471 0.3372549 0.3333333 0.3490196 0.3647059 0.3843138 0.3882353 0.3960785
 [2,] 0.2509804 0.3254902 0.4078432 0.4039216 0.4117647 0.4196079 0.4274510 0.4313726 0.4352942 0.4352942
 [3,] 0.1882353 0.1803922 0.2431373 0.2627451 0.2705882 0.2705882 0.2862745 0.2901961 0.2823530 0.2941177
 [4,] 0.2588235 0.3568628 0.4431373 0.4313726 0.4431373 0.4313726 0.4549020 0.4862745 0.5882353 0.6392157
 [5,] 0.2352941 0.2352941 0.2392157 0.2392157 0.2627451 0.2588235 0.2549020 0.2627451 0.2666667 0.2588235
 [6,] 0.2980392 0.3137255 0.3764706 0.4078432 0.4235294 0.4392157 0.4470589 0.4627451 0.4705883 0.4862745
 [7,] 0.2274510 0.2431373 0.2156863 0.2196079 0.2392157 0.2313726 0.2392157 0.2392157 0.2509804 0.2588235
 [8,] 0.3254902 0.3019608 0.3215686 0.4627451 0.5686275 0.5921569 0.6117647 0.6274510 0.6470588 0.6470588
 [9,] 0.3019608 0.2862745 0.2901961 0.3176471 0.3843138 0.3921569 0.3882353 0.4000000 0.4078432 0.4313726
[10,] 0.3215686 0.3215686 0.3254902 0.3098039 0.3058824 0.3176471 0.3215686 0.3294118 0.3372549 0.3411765
          [,11]     [,12]     [,13]     [,14]     [,15]     [,16]     [,17]     [,18]     [,19]     [,20]
 [1,] 0.4117647 0.4117647 0.3803922 0.2666667 0.3725490 0.3490196 0.3294118 0.3607843 0.3882353 0.3843138
 [2,] 0.4588236 0.4549020 0.4117647 0.4078432 0.4509804 0.4862745 0.4705883 0.4666667 0.4784314 0.4862745
 [3,] 0.3019608 0.3137255 0.3137255 0.2980392 0.1960784 0.2745098 0.3137255 0.3176471 0.3019608 0.3058824
 [4,] 0.6588235 0.6745098 0.6901961 0.5254902 0.6000000 0.6352941 0.6352941 0.6235294 0.6274510 0.6235294
 [5,] 0.2862745 0.2823530 0.2745098 0.2784314 0.2470588 0.2313726 0.2627451 0.2745098 0.2745098 0.2627451
 [6,] 0.5019608 0.5058824 0.5058824 0.4392157 0.3843138 0.4823530 0.5176471 0.5215687 0.5176471 0.5098040
 [7,] 0.2784314 0.2745098 0.2666667 0.2627451 0.2705882 0.2392157 0.2509804 0.2745098 0.2823530 0.2705882
 [8,] 0.6431373 0.6470588 0.6235294 0.5921569 0.6039216 0.5647059 0.5882353 0.6039216 0.6352941 0.6431373
 [9,] 0.4392157 0.4470589 0.4392157 0.4313726 0.4549020 0.4352942 0.4392157 0.4313726 0.4352942 0.4352942
[10,] 0.3215686 0.3254902 0.3333333 0.3215686 0.2980392 0.3294118 0.3607843 0.3725490 0.3607843 0.3686275

Answer 1

使用 matplotlib 你可以做到以下几点：

from matplotlib import pyplot as plt
plt.imshow(array_of_pixels)