import os import numpy as np from PIL import Image from tensorflow.keras.applications.resnet50 import ResNet50, preprocess_input from sklearn.cluster import KMeans # 加载 ResNet50 模型 model = ResNet50(weights='imagenet', include_top=False) # 加载并预处理图片，提取特征向量 features = [] for filename in os.listdir('D:/wjd/2'): if filename.endswith('.png'): img = Image.open(os.path.join('D:/wjd/2', filename)) img = img.convert('RGB') # 将 RGBA 转换为 RGB img = img.resize((224, 224)) x = np.array(img) x = np.expand_dims(x, axis=0) x = preprocess_input(x) feature = model.predict(x) feature = np.squeeze(feature) features.append(feature) # 将特征向量进行聚类，自动确定聚类的数量 kmeans = KMeans(n_clusters=None, max_iter=1000, n_init=50).fit(features) n_clusters = len(set(kmeans.labels_)) # 将每张图片分配到对应的聚类中 for i, label in enumerate(kmeans.labels_): filename = os.listdir('D:/wjd/2')[i] print('{} belongs to cluster {}'.format(filename, label))，可以将代码中的聚类方法换成 DBSCAN，其他保持不变

可以将代码中的聚类方法从KMeans换成DBSCAN，代码如下：

import os 
import numpy as np 
from PIL import Image 
from tensorflow.keras.applications.resnet50 import ResNet50, preprocess_input 
from sklearn.cluster import DBSCAN 

# 加载 ResNet50 模型 
model = ResNet50(weights='imagenet', include_top=False) 

# 加载并预处理图片，提取特征向量 
features = [] 
for filename in os.listdir('D:/wjd/2'): 
    if filename.endswith('.png'): 
        img = Image.open(os.path.join('D:/wjd/2', filename)) 
        img = img.convert('RGB') # 将 RGBA 转换为 RGB 
        img = img.resize((224, 224)) 
        x = np.array(img) 
        x = np.expand_dims(x, axis=0) 
        x = preprocess_input(x) 
        feature = model.predict(x) 
        feature = np.squeeze(feature) 
        features.append(feature) 

# 将特征向量进行聚类，自动确定聚类的数量 
dbscan = DBSCAN(eps=0.5, min_samples=2).fit(features) 
n_clusters = len(set(dbscan.labels_)) 

# 将每张图片分配到对应的聚类中 
for i, label in enumerate(dbscan.labels_): 
    filename = os.listdir('D:/wjd/2')[i] 
    print('{} belongs to cluster {}'.format(filename, label))