修改svm识别手写数字

Question

需要将提供的数据集和测试集用来调试，主要想采用（https://blog.csdn.net/weixin_41738030/article/details/100130638）这位楼主的实现方法，但如果要使用自己的测试集和数据集（都分为了标签和图片，格式是idx1-ubyte和idx3-ubyte），该如何修改？

Answer 1

你试试这个呢

import os
import cv2
import matplotlib.pyplot as plt
import numpy as np
from sklearn import datasets


def getdata(_dir):
    images= []
    labels = []
    for cls in os.listdir(_dir):
        cls_dir = os.path.join(_dir,cls)  
        for file in os.listdir(cls_dir ):
            img = cv2.imread(os.path.join(cls_dir,file ),-1)
            images.append(img)
            labels.append(int(cls))
    return images,labels


_dir_train=r""
_dir_test=r""
Xtrain,Ytrain = getdata(_dir_train)
Xtest,Ytest= getdata(_dir_test)

images_and_labels=list(zip(Xtrain,Ytrain))
plt.figure(figsize=(8, 6))
for index, (image, label) in enumerate(images_and_labels[:8]):
    plt.subplot(2, 4, index + 1)
    plt.axis('off')
    plt.imshow(image, cmap=plt.cm.gray_r, interpolation='nearest')
    plt.title('Digit: %i' % label, fontsize=20);

from sklearn import svm
clf = svm.SVC(gamma=0.001, C=100., probability=True)
clf.fit(Xtrain, Ytrain);

from sklearn.metrics import accuracy_score
Ypred = clf.predict(Xtest);
accuracy_score(Ytest, Ypred)

Answer 2

报错了怎么办

Answer 3

这回应该可以了

import os
import cv2
import matplotlib.pyplot as plt
import numpy as np
from sklearn import datasets

def getdata(_dir):
    images= []
    labels = []
    for cls in os.listdir(_dir):
        cls_dir = os.path.join(_dir,cls)  
        for file in os.listdir(cls_dir ):
            img = cv2.imread(os.path.join(cls_dir,file ),-1)
            images.append(img)
            labels.append(int(cls))
    return np.array(images),np.array(labels)


_dir_train=r"E:\test"
_dir_test=r"E:\test"
Xtrain,Ytrain = getdata(_dir_train)
Xtest,Ytest= getdata(_dir_test)

images_and_labels=list(zip(Xtrain,Ytrain))
plt.figure(figsize=(8, 6))
for index, (image, label) in enumerate(images_and_labels[:8]):
    plt.subplot(2, 4, index + 1)
    plt.axis('off')
    plt.imshow(image, cmap=plt.cm.gray_r, interpolation='nearest')
    plt.title('Digit: %i' % label, fontsize=20);

from sklearn import svm
clf = svm.SVC(gamma=0.001, C=100., probability=True)

h=Xtrain.shape[1]
w=Xtrain.shape[2]
Xtrain=Xtrain.reshape(-1,h*w)
Ytrain=Ytrain.reshape(-1,h*w)
clf.fit(Xtrain, Ytrain);

from sklearn.metrics import accuracy_score
Ypred = clf.predict(Xtest);
accuracy_score(Ytest, Ypred)

Answer 4

似乎还有一点点小问题