识别更多对象——多类分类

1# 导入所需库
2import numpy as np
3from PIL import Image
4from sklearn.model_selection import train_test_split
5from sklearn.svm import SVC
6from sklearn.metrics import accuracy_score
7
8def prepare_data():
9  # 使用Pillow读取图像文件
10  img_list = []
11  for i in range(1, 301):
12      img = Image.open('./gray_g'+str(i)+'.png').convert('L')
13      img_list.append(img)
14  
15  # 将图像数据重塑为一维数组
16  img_data = np.array([np.array(img).flatten() for img in img_list])
17  
18  # 前100张图像标签为0，接下来的100张标签为1，最后100张标签为2
19  tag = np.concatenate((np.zeros(100), np.ones(100), np.ones(100)*2))
20  
21  return img_data, tag
22
23def get_model():
24  # 返回选定的模型
25  return SVC()
26  
27def train_model(model, data, tags):
28  
29  # 训练支持向量机分类器
30  model.fit(data, tags)
31  
32  return model
33  
34def get_predictions(model, test_img):
35  
36  # 获取预测结果
37  predictions = model.predict(test_img)
38  
39  return predictions
40  
41def evaluate_predictions(predicted_tags, actual_tags):
42  
43  # 计算预测准确率
44  accuracy = accuracy_score(actual_tags, predicted_tags)
45  
46  print("预测准确率:", accuracy)
47
48
49def main():
50
51  # 数据准备
52  img_data, tags = prepare_data()
53  
54  # 将数据划分为训练集和测试集
55  train_img, test_img, train_tags, test_tags = train_test_split(img_data, tags, test_size=0.2)
56    
57  # 模型初始化
58  model = get_model()
59  
60  # 模型训练
61  trained_model = train_model(model, train_img, train_tags)
62  
63  # 预测阶段
64  predictions = get_predictions(trained_model, test_img)
65  
66  # 评估预测结果
67  evaluate_predictions(predictions, test_tags)
68  
69main()