如何使用自己的图像使用 MNIST 数据集和神经网络预测手写数字

Question

我已经使用神经网络构建了一个模型，该模型使用 MNIST 数据集来预测数字。

现在我想对我提供给模型的图像进行预测。

我正在使用此代码将图像转换为 MNIST 图像。

from PIL import Image, ImageFilter
from matplotlib import pyplot as plt


def imageprepare(argv):
    """
    This function returns the pixel values.
    The imput is a png file location.
    """
    im = Image.open(argv).convert('L')
    width = float(im.size[0])
    height = float(im.size[1])
    newImage = Image.new('L', (28, 28), (255))  # creates white canvas of 28x28 pixels

    if width > height:  # check which dimension is bigger
        # Width is bigger. Width becomes 20 pixels.
        nheight = int(round((20.0 / width * height), 0))  # resize height according to ratio width
        if (nheight == 0):  # rare case but minimum is 1 pixel
            nheight = 1
            # resize and sharpen
        img = im.resize((20, nheight), Image.ANTIALIAS).filter(ImageFilter.SHARPEN)
        wtop = int(round(((28 - nheight) / 2), 0))  # calculate horizontal position
        newImage.paste(img, (4, wtop))  # paste resized image on white canvas
    else:
        # Height is bigger. Heigth becomes 20 pixels.
        nwidth = int(round((20.0 / height * width), 0))  # resize width according to ratio height
        if (nwidth == 0):  # rare case but minimum is 1 pixel
            nwidth = 1
            # resize and sharpen
        img = im.resize((nwidth, 20), Image.ANTIALIAS).filter(ImageFilter.SHARPEN)
        wleft = int(round(((28 - nwidth) / 2), 0))  # caculate vertical pozition
        newImage.paste(img, (wleft, 4))  # paste resized image on white canvas

    # newImage.save("sample.png

    tv = list(newImage.getdata())  # get pixel values

    # normalize pixels to 0 and 1. 0 is pure white, 1 is pure black.
    tva = [(255 - x) * 1.0 / 255.0 for x in tv]
    print(tva)
    return tva

x=[imageprepare('./zero.png')]#file path here
print(len(x))# mnist IMAGES are 28x28=784 pixels
print(x[0])
#Now we convert 784 sized 1d array to 24x24 sized 2d array so that we can visualize it
newArr=[[0 for d in range(28)] for y in range(28)]
k = 0
for i in range(28):
    for j in range(28):
        newArr[i][j]=x[0][k]
        k=k+1

for i in range(28):
    for j in range(28):
        print(newArr[i][j])
        # print(' , ')
    print('\n')


plt.imshow(newArr, interpolation='nearest')
plt.savefig('MNIST_IMAGE2.png')#save MNIST image
plt.show()#Show / plot that image

我正在使用以下代码来预测结果：

image = cv2.imread("MNIST_IMAGE2.png")
image = cv2.resize(image,(28,28))
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
data = asarray(gray)
data=data/255.0
predictions = model.predict(np.expand_dims(data, 0))

但是使用上述代码生成的输出非常不准确。

我该怎么办？？

Answer 1

如果您想要更准确地识别您书写的数字，您可能需要使用您书写的数字图像而不是 mnist 的图像来训练网络。