深层神经网络——小猫分类器

1 问题背景

不借助工具包，从0开始搭建一个5层或多层的二类分类器，也就是对图片中是否有小猫进行分类。问题输入为64*64的RGB图像，当然也可以改变输入大小，有2个途径：一个改变图片大小，二是改变输入层大小。问题输入为0或1，当前也可以改变输出层大小实现更多目标分类。

2 环境设置

import numpy as np    #基本操作
import h5py    #数据集
import matplotlib.pyplot as plt    #画图
from PIL import Image        #图像处理
from dnn_app_utils_v2 import * #

%matplotlib inline #画图的设置
plt.rcParams['figure.figsize'] = (5.0, 4.0) # set default size of plots
plt.rcParams['image.interpolation'] = 'nearest'
plt.rcParams['image.cmap'] = 'gray'

3 数据处理

输入是一张图片，大小为64*64*3，值范围为（0,255），需要处理为12288*1，值范围（0,1）。

#数据导入
train_x_orig, train_y, test_x_orig, test_y, classes = load_data()
#数据大小和数量
m_train = train_x_orig.shape[0]    #209
num_px = train_x_orig.shape[1]    #64
m_test = test_x_orig.shape[0]    #50
#
train_x_flatten = train_x_orig.reshape(train_x_orig.shape[0], -1).T  
test_x_flatten = test_x_orig.reshape(test_x_orig.shape[0], -1).T
train_x = train_x_flatten/255.     #shape(12288, 209)
test_x = test_x_flatten/255.    #shape(12288, 50)

4 模型框架

#模型层数
layers_dims = [12288, 20, 7, 5, 1] 
#循环训练
def L_layer_model(X, Y, layers_dims, learning_rate = 0.0075, num_iterations = 3000, print_cost=False):
    np.random.seed(1)
    costs = []    # 损失记录
    #初始化W、b
    parameters = initialize_parameters_deep(layers_dims)
    #迭代训练
    for i in range(0, num_iterations):
        #前向传播
        AL, caches = L_model_forward(X, parameters)
        #损失计算
        cost = compute_cost(AL, Y)
        #反向传播
        grads = L_model_backward(AL, Y, caches)
        #更新W、b
        parameters = update_parameters(parameters, grads, learning_rate)
        # 损失值观测
        if print_cost and i % 100 == 0:
            print ("Cost after iteration %i: %f" %(i, cost))
        if print_cost and i % 100 == 0:
            costs.append(cost)
    # 损失曲线
    plt.plot(np.squeeze(costs))
    plt.ylabel('cost')
    plt.xlabel('iterations (per tens)')
    plt.title("Learning rate =" + str(learning_rate))
    plt.show()
    return parameters

5  模型训练

parameters = L_layer_model(train_x, train_y, layers_dims, num_iterations = 2500, print_cost = True)

Cost after iteration 0: 0.771749
Cost after iteration 100: 0.672053
Cost after iteration 200: 0.648263
Cost after iteration 300: 0.611507
Cost after iteration 400: 0.567047
Cost after iteration 500: 0.540138
Cost after iteration 600: 0.527930
Cost after iteration 700: 0.465477
Cost after iteration 800: 0.369126
Cost after iteration 900: 0.391747
Cost after iteration 1000: 0.315187
Cost after iteration 1100: 0.272700
Cost after iteration 1200: 0.237419
Cost after iteration 1300: 0.199601
Cost after iteration 1400: 0.189263
Cost after iteration 1500: 0.161189
Cost after iteration 1600: 0.148214
Cost after iteration 1700: 0.137775
Cost after iteration 1800: 0.129740
Cost after iteration 1900: 0.121225
Cost after iteration 2000: 0.113821
Cost after iteration 2100: 0.107839
Cost after iteration 2200: 0.102855
Cost after iteration 2300: 0.100897
Cost after iteration 2400: 0.092878

6 模型预测

pred_train = predict(train_x, train_y, parameters)
pred_test = predict(test_x, test_y, parameters)

Accuracy: 0.9856459330143539
Accuracy: 0.8

7 误差分析

print_mislabeled_images(classes, test_x, test_y, pred_test)

• 猫的身体处于不寻常的位置‎
• ‎猫出现在颜色相似的背景上‎
• ‎不寻常的猫的颜色和种类‎
• ‎相机角度‎
• ‎图片亮度‎
• ‎比例变化（猫在图像中非常大或很小） 

8 测试自己的图片

用训练好的权值参数