基于卷积神经网络的手写数字识别（非mnist数据集）

1. 前言

前不久在学习机器学习，下载了一些手写数字，sklearn没有很好的利用起来，这几天学习tensorflow，但一直想用的这些手写数字做训练，因此用CNN写了个简单的手写数字识别，正确率可达90%以上，图片代码数据可点我下载。

2. 代码部分

手写数字加载代码：

import cv2.cv2 as cv
import glob
import numpy as np
from sklearn.model_selection import train_test_split
import os
def load_pic(img_path, num):
    for step, pic_path in enumerate(glob.glob(img_path)):
        pic = cv.imread(pic_path)
        # pic = cv.cvtColor(pic,cv.COLOR_RGB2GRAY)
        pic = cv.resize(pic,(64, 64),interpolation=cv.INTER_CUBIC)
        pic = pic[np.newaxis, ...]

        if step == 0:
            data = pic
            label_num = [num]
        else:
            data = np.concatenate((data, pic), axis=0)
            label_num.append(num)
    return data, label_num

def load_num():
    for num in range(10):

        path_1 = os.getcwd()
        path_2 = r'\*.bmp'
        path = path_1 + '\handwritten' + '\\' + str(num) + path_2
        number, label = load_pic(path, num)
        label = np.array(label)
        if num == 0:
            data = number
            label_all = label
        else:
            data = np.concatenate((data, number), axis=0)
            label_all = np.hstack((label_all, label))
    return data, label_all
def load_data(size = 0.2):
    data, label = load_num()
    X_train, X_test, y_train, y_test = train_test_split(data, label, test_size=size, random_state=42) # 数据按%80，%20划分
    return X_train, X_test, y_train, y_test

卷积神经网络代码：

import os
import load_pic as lp
os.environ['TF_CPP_MIN_LOG_LEVEL']='2'

import  tensorflow as tf
from    tensorflow.keras import layers, optimizers, datasets, Sequential
from tensorflow.compat.v1 import ConfigProto
from tensorflow.compat.v1 import InteractiveSession

config = ConfigProto()
config.gpu_options.allow_growth = True
session = InteractiveSession(config=config)

tf.random.set_seed(2345)


conv_layers = [ # 5 units of conv + max pooling
    # unit 1
    layers.Conv2D(64, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(64, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    # unit 2
    layers.Conv2D(128, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(128, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    # unit 3
    layers.Conv2D(256, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(256, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    # unit 4
    layers.Conv2D(256, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(256, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    # unit 5
    layers.Conv2D(512, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(512, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    # unit 6
    layers.Conv2D(512, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(512, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same')
]
def preprocess(x, y):
    # [0~1]
    x = tf.cast(x, dtype=tf.float32) / 255.
    y = tf.cast(y, dtype=tf.int32)
    return x,y
X_train, X_test, y_train, y_test = lp.load_data()
print(X_train.shape, y_train.shape, X_test.shape, y_test.shape)
train_db = tf.data.Dataset.from_tensor_slices((X_train,y_train))
train_db = train_db.shuffle(500).map(preprocess).batch(128)
test_db = tf.data.Dataset.from_tensor_slices((X_test,y_test))
test_db = test_db.map(preprocess).batch(64)

sample = next(iter(train_db))
print('sample:', sample[0].shape, sample[1].shape,
      tf.reduce_min(sample[0]), tf.reduce_max(sample[0]))

def main():

    conv_net = Sequential(conv_layers)

    fc_net = Sequential([
        layers.Dense(256, activation=tf.nn.relu),
        layers.Dense(64, activation=tf.nn.relu),
        layers.Dense(10, activation=None),
    ])

    conv_net.build(input_shape=[None, 75, 75, 3])
    fc_net.build(input_shape=[None, 512])
    optimizer = optimizers.Adam(lr=1e-4)

    # [1, 2] + [3, 4] => [1, 2, 3, 4]
    variables = conv_net.trainable_variables + fc_net.trainable_variables

    for epoch in range(50):

        for step, (x,y) in enumerate(train_db):

            with tf.GradientTape() as tape:
                # print(x.shape)
                out = conv_net(x)
                out = tf.reshape(out, [-1, 512])
                # print(out.shape)
                logits = fc_net(out)
                # [b] => [b, 10]
                y_onehot = tf.one_hot(y, depth=10)
                # compute loss
                # print(logits.shape)
                # print(y_onehot.shape)
                loss = tf.losses.categorical_crossentropy(y_onehot, logits, from_logits=True)
                loss = tf.reduce_mean(loss)

            grads = tape.gradient(loss, variables)
            optimizer.apply_gradients(zip(grads, variables))

            if step %3 == 0:
                print(epoch, step, 'loss:', float(loss))
        total_num = 0
        total_correct = 0
        for x,y in test_db:

            out = conv_net(x)
            out = tf.reshape(out, [-1, 512])
            logits = fc_net(out)
            prob = tf.nn.softmax(logits, axis=1)
            pred = tf.argmax(prob, axis=1)
            pred = tf.cast(pred, dtype=tf.int32)

            correct = tf.cast(tf.equal(pred, y), dtype=tf.int32)
            correct = tf.reduce_sum(correct)

            total_num += x.shape[0]
            total_correct += int(correct)

        acc = total_correct / total_num
        print(epoch, 'acc:', acc)


if __name__ == '__main__':
    main()