目录
Python手写字母识别
准备
#设置镜像
pip install pip -U
pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
#安装
pip install numpy
pip install pandas
pip install scipy
pip install scikit-learn
pip install imageio
数据下载
Chars74K数据集：链接
使用EnglishHnd数据集，每个字母/数字有55张
由于原始数据的格式不方便处理，将其处理为以下格式，然后把每张图片缩小为40x30：
img
--- 0
------ 0.jpg
------ 1.jpg
------ ...
------ 54.jpg
--- 1
------ ...
--- lower_a
------ ...
--- ...
--- upper_a
------ ...
--- ...
处理过程我就不细说了，可以直接下载处理后的数据集
数据预处理
将数据集处理为CSV格式：
import numpy as np
import imageio
import os
#将各个数字、字母对应0-61中的一个数字
classes=os.listdir('img')
f=open("map.csv","w")
for i,j in enumerate(classes):
print(str(i)+','+j,file=f)
#用imageio库读入图片
x=[]
y=[]
for i in classes:
for j in range(55):
img=imageio.imread("img/{}/{}.jpg".format(i,j))
bw=[]
for k in img:
bw.append(l[0])
x_train.append(bw)
y_train.append(classes.index(i))
#导出csv
f=open("chars.csv","w")
for i,j in zip(x,y):
print(','.join(i)+','+str(j),file=f)
得到map.csv：
0,0
......
10,lower_a
......
36,upper_a
......
chars.csv:
0,0,...,0,0,0
......
0,0,...,0,0,61
训练
首先导入包：
from pandas import read_csv
from pandas.plotting import scatter_matrix
from matplotlib import pyplot
from sklearn.model_selection import train_test_split
from sklearn.model_selection import KFold
from sklearn.model_selection import cross_val_score
from sklearn.metrics import classification_report
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
from sklearn.linear_model import LogisticRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.svm import SVC
import joblib
评估模型：
#读取、处理
filename = 'chars.csv'
dataset = read_csv(filename)
array = dataset.values
X = array[:, 0:784]
Y = array[:, 784]
validation_size = 0.2
seed = 7
X_train, X_validation, Y_train, Y_validation = \
train_test_split(X, Y, test_size=validation_size, random_state=seed)
#模型选择
models = {}
models['LR'] = LogisticRegression()
models['LDA'] = LinearDiscriminantAnalysis()
models['KNN'] = KNeighborsClassifier()
models['CART'] = DecisionTreeClassifier()
models['NB'] = GaussianNB()
models['SVM'] = SVC()
results = []
for key in models:
kfold = KFold(n_splits=10, random_state=seed, shuffle=True)
cv_results = cross_val_score(models[key], X_train, Y_train, cv=kfold, scoring='accuracy')
results.append(cv_results)
print('%s: %f (%f)' %(key, cv_results.mean(), cv_results.std()))
#可视化
fig = pyplot.figure()
fig.suptitle('Algorithm Comparison')
ax = fig.add_subplot(111)
pyplot.boxplot(results)
ax.set_xticklabels(models.keys())
pyplot.show()
结果如下：
LR: 0.916341 (0.017975)
LDA: 0.906424 (0.013931)
KNN: 0.838157 (0.013883)
CART: 0.893944 (0.023094)
NB: 0.235954 (0.023780)
SVM: 0.916711 (0.017377)
可以看出SVM效果最好，使用SVM分类（SVM训练时间可能较长）：
svm=SVC()
svm.fit(X=X_train, y=Y_train)
predictions = svm.predict(X_validation)
print(accuracy_score(Y_validation, predictions))
print(confusion_matrix(Y_validation, predictions))
print(classification_report(Y_validation, predictions))
joblib.dump(svm,"letters.model")
0.9046920821114369
[[502 29]
[ 36 115]]
precision recall f1-score support
0 0.93 0.95 0.94 531
1 0.80 0.76 0.78 151
accuracy 0.90 682
macro avg 0.87 0.85 0.86 682
weighted avg 0.90 0.90 0.90 682
准确率在90%左右。由于大小写是分离的，做到这个准确率已经很不错了。如果应用不需要区分大小写，可以将同一个字母的大小写放在一起，大家可以尝试一下。
这段程序导出了一个模型。下次使用时，可以使用与处理训练图片相同的方法进行图片处理，然后使用模型预测。
有任何不足之处，欢迎评论区留言