改进使用 mnist 数据集训练的神经网络的真实结果

Improve real-life results of neural network trained with mnist dataset

我已经使用 mnist 数据集用 keras 构建了一个神经网络,现在我正尝试将它用于实际手写数字的照片。当然我并不期望结果是完美的,但我目前得到的结果还有很大的改进空间。

对于初学者,我使用一些以我最清晰的笔迹书写的单个数字的照片对其进行测试。它们是正方形的,并且与 mnist 数据集中的图像具有相同的尺寸和颜色。它们保存在名为 individual_test 的文件夹中,例如:7(2)_digit.jpg.

网络通常非常确定错误的结果,我会给你举个例子:

这张图片我得到的结果如下:

result:  3 . probabilities:  [1.9963557196245318e-10, 7.241294497362105e-07, 0.02658148668706417, 0.9726449251174927, 2.5416460047722467e-08, 2.6078915027483163e-08, 0.00019745019380934536, 4.8302300825753264e-08, 0.0005754049634560943, 2.8358477788259506e-09]

所以网络有 97% 的把握确定这是一个 3,而这张照片并不是唯一的例子。在 38 张图片中,只有 16 张被正确识别。令我震惊的是,网络对它的结果如此确定,尽管它与正确结果相去甚远。

编辑
将阈值添加到 prepare_image (img = cv2.threshold(img, 0.1, 1, cv2.THRESH_BINARY_INV)[1]) 后,性能略有提高。现在 38 张图片中有 19 张是正确的,但对于一些图片,包括上面显示的图片,它仍然非常确定错误的结果。这是我现在得到的:

result:  3 . probabilities:  [1.0909866760000497e-11, 1.1584616004256532e-06, 0.27739930152893066, 0.7221096158027649, 1.900260038212309e-08, 6.555900711191498e-08, 4.479645940591581e-05, 6.455550760620099e-07, 0.0004443934594746679, 1.0013242457418414e-09]

所以现在只有 72% 确定它的结果哪个更好但仍然...



我可以做些什么来提高性能?我可以更好地准备图像吗?或者我应该将自己的图像添加到训练数据中吗?如果是这样,我会怎么做?

编辑

上面显示的图片在应用 prepare_image 后的样子:

使用阈值后,这是同一张图片的样子:

对比一下:这是mnist数据集提供的其中一张图片:

他们看起来和我很相似。我该如何改进呢?
这是我的代码(包括阈值):

# import keras and the MNIST dataset
from tensorflow.keras.datasets import mnist
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from keras.utils import np_utils
# numpy is necessary since keras uses numpy arrays
import numpy as np

# imports for pictures
import matplotlib.pyplot as plt
import PIL
import cv2

# imports for tests
import random
import os

class mnist_network():
    def __init__(self):
        """ load data, create and train model """
        # load data
        (X_train, y_train), (X_test, y_test) = mnist.load_data()
        # flatten 28*28 images to a 784 vector for each image
        num_pixels = X_train.shape[1] * X_train.shape[2]
        X_train = X_train.reshape((X_train.shape[0], num_pixels)).astype('float32')
        X_test = X_test.reshape((X_test.shape[0], num_pixels)).astype('float32')
        # normalize inputs from 0-255 to 0-1
        X_train = X_train / 255
        X_test = X_test / 255
        # one hot encode outputs
        y_train = np_utils.to_categorical(y_train)
        y_test = np_utils.to_categorical(y_test)
        num_classes = y_test.shape[1]


        # create model
        self.model = Sequential()
        self.model.add(Dense(num_pixels, input_dim=num_pixels, kernel_initializer='normal', activation='relu'))
        self.model.add(Dense(num_classes, kernel_initializer='normal', activation='softmax'))
        # Compile model
        self.model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

        # train the model
        self.model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=10, batch_size=200, verbose=2)

        self.train_img = X_train
        self.train_res = y_train
        self.test_img = X_test
        self.test_res = y_test


    def predict_result(self, img, show = False):
        """ predicts the number in a picture (vector) """
        assert type(img) == np.ndarray and img.shape == (784,)

        if show:
            img = img.reshape((28, 28))
            # show the picture
            plt.imshow(img, cmap='Greys')
            plt.show()
            img = img.reshape(img.shape[0] * img.shape[1])

        num_pixels = img.shape[0]
        # the actual number
        res_number = np.argmax(self.model.predict(img.reshape(-1,num_pixels)), axis = 1)
        # the probabilities
        res_probabilities = self.model.predict(img.reshape(-1,num_pixels))

        return (res_number[0], res_probabilities.tolist()[0])    # we only need the first element since they only have one


    def prepare_image(self, img, show = False):
        """ prepares the partial images used in partial_img_rec by transforming them
            into numpy arrays that the network will be able to process """
        # convert to greyscale
        img = img.convert("L")
        # rescale image to 28 *28 dimension
        img = img.resize((28,28), PIL.Image.ANTIALIAS)
        # inverse colors since the training images have a black background
        #img =  PIL.ImageOps.invert(img)
        # transform to vector
        img = np.asarray(img, "float32")
        img = img / 255.
        img[img < 0.5] = 0.

        img = cv2.threshold(img, 0.1, 1, cv2.THRESH_BINARY_INV)[1]

        if show:
            plt.imshow(img, cmap = "Greys")

        # flatten image to 28*28 = 784 vector
        num_pixels = img.shape[0] * img.shape[1]
        img = img.reshape(num_pixels)

        return img


    def partial_img_rec(self, image, upper_left, lower_right, results=[], show = False):
        """ partial is a part of an image """
        left_x, left_y = upper_left
        right_x, right_y = lower_right

        print("current test part: ", upper_left, lower_right)
        print("results: ", results)
        # condition to stop recursion: we've reached the full width of the picture
        width, height = image.size
        if right_x > width:
            return results

        partial = image.crop((left_x, left_y, right_x, right_y))
        if show:
            partial.show()
        partial = self.prepare_image(partial)

        step = height // 10

        # is there a number in this part of the image? 
        res, prop = self.predict_result(partial)
        print("result: ", res, ". probabilities: ", prop)
        # only count this result if the network is at least 50% sure
        if prop[res] >= 0.5:        
            results.append(res)
            # step is 80% of the partial image's size (which is equivalent to the original image's height) 
            step = int(height * 0.8)
            print("found valid result")
        else:
            # if there is no number found we take smaller steps
            step = height // 20 
        print("step: ", step)
        # recursive call with modified positions ( move on step variables )
        return self.partial_img_rec(image, (left_x + step, left_y), (right_x + step, right_y), results = results)

    def individual_digits(self, img):
        """ uses partial_img_rec to predict individual digits in square images """
        assert type(img) == PIL.JpegImagePlugin.JpegImageFile or type(img) == PIL.PngImagePlugin.PngImageFile or type(img) == PIL.Image.Image

        return self.partial_img_rec(img, (0,0), (img.size[0], img.size[1]), results=[])

    def test_individual_digits(self):
        """ test partial_img_rec with some individual digits (shape: square) 
            saved in the folder 'individual_test' following the pattern 'number_digit.jpg' """
        cnt_right, cnt_wrong = 0,0
        folder_content = os.listdir(".\individual_test")

        for imageName in folder_content:
            # image file must be a jpg or png
            assert imageName[-4:] == ".jpg" or imageName[-4:] == ".png"
            correct_res = int(imageName[0])
            image = PIL.Image.open(".\individual_test\" + imageName).convert("L")
            # only square images in this test
            if image.size[0]  != image.size[1]:
                print(imageName, " has the wrong proportions: ", image.size,". It has to be a square.")
                continue 
            predicted_res = self.individual_digits(image)

            if predicted_res == []:
                print("No prediction possible for ", imageName)
            else:
                predicted_res = predicted_res[0]

            if predicted_res != correct_res:
                print("error in partial_img-rec! Predicted ", predicted_res, ". The correct result would have been ", correct_res)
                cnt_wrong += 1
            else:
                cnt_right += 1
                print("correctly predicted ",imageName)
        print(cnt_right, " out of ", cnt_right + cnt_wrong," digits were correctly recognised. The success rate is therefore ", (cnt_right / (cnt_right + cnt_wrong)) * 100," %.")

    def multiple_digits(self, img):
        """ takes as input an image without unnecessary whitespace surrounding the digits """

        #assert type(img) == myImage
        width, height = img.size
        # start with the first square part of the image
        res_list = self.partial_img_rec(img, (0,0),(height ,height), results = [])
        res_str = ""
        for elem in res_list:
            res_str += str(elem)
        return res_str

    def test_multiple_digits(self):
        """ tests the function 'multiple_digits' using some images saved in the folder 'multi_test'.
            These images contain multiple handwritten digits without much whitespac surrounding them.
            The correct solutions are saved in the files' names followed by the characte '_'. """

        cnt_right, cnt_wrong = 0,0
        folder_content = os.listdir(".\multi_test")
        for imageName in folder_content:
            # image file must be a jpg or png
            assert imageName[-4:] == ".jpg" or imageName[-4:] == ".png"            
            image = PIL.Image.open(".\multi_test\" + imageName).convert("L")

            correct_res = imageName.split("_")[0]
            predicted_res = self.multiple_digits(image)
            if correct_res == predicted_res:
                cnt_right += 1
            else:
                cnt_wrong += 1
                print("Error in multiple_digits! The network predicted ", predicted_res, " but the correct result would have been ", correct_res)

        print("The network predicted correctly ", cnt_right, " out of ", cnt_right + cnt_wrong, " pictures. That's a success rate of ", cnt_right / (cnt_right + cnt_wrong) * 100, "%.")

network = mnist_network()
# this is the image shown above
result = network.individual_digits(PIL.Image.open(".\individual_test\7(2)_digit.jpg"))

你在 MNIST 数据集上的测试成绩是多少? 我想到的一件事是您的图像缺少阈值,

阈值化是一种将低于某个像素的像素值设为零的技术,请参阅任何地方的 OpenCV 阈值化示例,您可能需要使用逆阈值化并再次检查您的结果。

做,有进展通知

你遇到的主要问题是你正在测试的图像与 MNIST 图像不同,可能是由于你已经完成了图像的准备工作,你能展示你正在测试的图像吗?在上面应用 prepare_image。

更新:

您有三种选择可以在此特定任务中获得更好的表现:

  1. 使用卷积网络,因为它在具有空间数据的任务中表现更好,例如图像,并且是更具生成性的分类器,例如这个。
  2. 使用或创建 and/or 生成更多您的类型的图片训练您的网络 使您的网络能够也要学习它们。
  3. 预处理您的图像以更好地与原始 MNIST 图像对齐,您之前曾针对这些图像训练过您的网络。

我刚刚做了一个实验。我检查了关于每个代表数字的 MNIST 图像。我拍摄了您的图像并进行了一些我之前向您建议的预处理,例如:

1.做了一些阈值,只是向下消除了背景噪声,因为原始MNIST数据有一些最小阈值只针对空白背景:

image[image < 0.1] = 0.

2. 令人惊讶的是,图像内部数字的大小被证明是至关重要的,所以我缩放了 28 x 28 图像内部的数字,例如我们在数字周围有更多的填充。

3. 我反转了图像,因为来自 keras 的 MNIST 数据也反转了。

image = ImageOps.invert(image)

4. 最后缩放数据,正如我们在训练中所做的那样:

image = image / 255.

预处理后,我使用参数 epochs=12, batch_size=200 的 MNIST 数据集训练模型,结果:

结果:1 概率:0.6844741106033325

 result:  **1** . probabilities:  [2.0584749904628552e-07, 0.9875971674919128, 5.821426839247579e-06, 4.979299319529673e-07, 0.012240586802363396, 1.1566483948399764e-07, 2.382085284580171e-08, 0.00013023221981711686, 9.620113416985987e-08, 2.5273093342548236e-05]

结果:6 概率:0.9221984148025513

result:  6 . probabilities:  [9.130864782491699e-05, 1.8290626258021803e-07, 0.00020504613348748535, 2.1564576968557958e-07, 0.0002401985548203811, 0.04510130733251572, 0.9221984148025513, 1.9014490248991933e-07, 0.03216308355331421, 3.323434683011328e-08]

结果:7 概率:0.7105212807655334 注:

result:  7 . probabilities:  [1.0372193770535887e-08, 7.988557626958936e-06, 0.00031014863634482026, 0.0056108818389475346, 2.434678014751057e-09, 3.2280522077599016e-07, 1.4190952857262573e-09, 0.9940618872642517, 1.612859932720312e-06, 7.102244126144797e-06]

您的号码 9 有点棘手:

当我弄清楚 MNIST 数据集的模型时,关于 9 的两个主要 "features"。上部和下部。与您的图像一样,具有漂亮圆形的上半部分不是 9,但对于您针对 MNIST 数据集训练的模型,主要是 3。根据 MNIST 数据集,9 的下半部分大部分是拉直曲线。所以基本上你的完美形状 9 对于你的模型总是 3 因为 MNIST 样本,除非你再次训练模型有足够的数量您形状的样本 9。为了验证我的想法,我用 9s:

做了一个子实验

我的 9 上部倾斜(根据 MNIST,对于 9 大部分都可以)但底部略微卷曲(不适用于9 根据 MNIST):

结果:9 概率:0.5365301370620728

我的 9 上部倾斜(根据 MNIST,对于 9 大部分都可以)并且底部是直的(对于 9 根据 MNIST):

结果:9 概率:0.923724353313446

你的 9 具有被误解的形状属性:

结果:3 概率:0.8158268928527832

result:  3 . probabilities:  [9.367801249027252e-05, 3.9978775021154433e-05, 0.0001467708352720365, 0.8158268928527832, 0.0005801069783046842, 0.04391581565141678, 6.44062723154093e-08, 7.099170943547506e-06, 0.09051419794559479, 0.048875387758016586]


最后证明图像缩放(填充)的重要性,我在上面提到的至关重要:

结果:3 概率:0.9845736622810364

结果:9 概率:0.923724353313446

所以我们可以看到我们的模型拾取了一些特征,它解释,在图像内部的超大形状和低填充尺寸的情况下总是分类为 3 .

我认为我们可以通过 CNN 获得更好的性能,但采样和预处理的方式对于在 ML 任务中获得最佳性能始终至关重要。

希望对您有所帮助。

更新二:

我发现了另一个问题,我也检查过并证明是正确的,图像中数字的位置也很重要,这对于这种类型的神经网络来说是有意义的。一个很好的例子,数字 79 被放置在 MNIST 数据集中的中心,靠近图像的底部导致更难或 flase 分类如果我们将用于分类的新数字放在图像的中心。我检查了将 7s 和 9s 移向底部的理论,因此在图像顶部留下更多位置,结果是几乎 100% 准确度。 由于这是一个 spatial 类型的问题,我想,使用 CNN 我们可以更有效地消除它。然而,如果 MNIST 对齐到中心会更好,或者我们可以通过编程来避免这个问题。