【机器学习】六种算法在人脸补全中的应用比较（K紧邻 线性 决策树 岭回归 套索回

需求：

根据人的上半边脸预测下半边脸，用各种算法取得的结果与原图比较

思考：

这是一个回归问题，不是分类问题（人脸数据不固定） 数据集一共包含40个人，每一个人10张照片，分布规律

每一个人取出8张照片作为训练数据，2张照片作为测试数据 样本特征和样本标签如何拆分？上半边脸作为样本特征，下半边脸作为特征标签

————————————————

人脸图像补全的方法用途及研究

导包

import numpy as npimport matplotlib.pyplot as plt%matplotlib inline#构建方程from sklearn.linear_model import LinearRegression,Ridge,Lasso#不会构建方程from sklearn.neighbors import KNeighborsRegressorfrom sklearn.tree import DecisionTreeRegressorfrom sklearn import datasetsfrom sklearn.model_selection import train_test_split

导入datasets中400位人脸数据数据

faces = datasets.fetch_olivetti_faces()X = faces.dataimages = faces.imagesy = faces.targetdisplay(X.shape)display(images.shape)display(y.shape)

(400, 4096)

(400, 64, 64)

(400,)

plt.figure(figsize=(2,2))index = np.random.randint(0,400,size =1)[0]img = images[index]plt.imshow(img,cmap = plt.cm.gist_gray)

将X(人脸数据)分成上半张人脸和下半张人脸

X_up = X[:,:2048]X_down = X[:,2048:] index = np.random.randint(0,400,size =1)[0]axes = plt.subplot(1,3,1)up_face = X_up[index].reshape(32,64)axes.imshow(up_face,cmap = plt.cm.gray)axes = plt.subplot(1,3,2)down_face = X_down[index].reshape(32,64)axes.imshow(down_face,cmap = plt.cm.gray)axes = plt.subplot(1,3,3)face = X[index].reshape(64,64)axes.imshow(face,cmap = plt.cm.gray)

X = X_up.copy()y = X_down.copy()display(X.shape,y.shape)

(400, 2048)

(400, 2048)

32*64

2048

X_train,X_test,y_train,y_test = train_test_split(X,y,test_size =30) estimators = {}#线性回归estimators['linear'] = LinearRegression()estimators['ridge'] = Ridge(alpha=0.1)estimators['knn'] = KNeighborsRegressor(n_neighbors=5)estimators['lasso'] = Lasso(alpha=0.1)estimators['ElasticNet'] = ElasticNet()estimators['tree'] = DecisionTreeRegressor()#决策树费时间 2048个样本特征#criterion = 'mse' 线性的是gini 和熵 都是越小越好

分别调用这六个每个算法

result = {}for key,model in estimators.items():model.fit(X_train,y_train)y_ = model.predict(X_test)#预测的是下班长人脸result[key] = y_

结果可视化

plt.figure(figsize=(8*2,2*10,))for i in range(0,10):#绘制第一列，上班张人脸axes = plt.subplot(10,8,i*8+1)up_face = X_test[i].reshape(32,64)axes.imshow(up_face,cmap= plt.cm.gray)#取消刻度axes.axis('off')#设置标题(只在第一列显示)if i == 0:axes.set_title('upface')#第七列绘制整张人脸axes = plt.subplot(10,8,i*8+8)down_face = y_test[i].reshape(32,64)#上下脸拼接true_face = np.concatenate([up_face,down_face])axes.imshow(true_face,cmap= plt.cm.gray) axes.axis('off')if i == 0:axes.set_title('trueface')#绘制第二列到第六列 ，算法预测的数据result,#字典 key 算法value 预测人脸#用enumerate 循环增加了个jfor j , key in enumerate(result): #j,0,1,2,3,4axes = plt.subplot(10,8,i*8+2+j)y_ = result[key]pre_downface = y_[i].reshape(32,64)pre_face = np.concatenate([up_face,pre_downface])axes.imshow(pre_face,cmap = plt.cm.gray)axes.axis('off')if i == 0:axes.set_title(key)

【机器学习】六种算法在人脸补全中的应用比较（K紧邻 线性 决策树 岭回归 套索回归 ElasticNet）