11.11图片智能识别推荐

import tensorflow as tf
from tensorflow.keras.applications import VGG16
from sklearn.metrics.pairwise import cosine_similarity
import numpy as np

# 加载预训练的VGG16模型
model = VGG16(weights='imagenet', include_top=False)

def extract_features(image_path):
    img = tf.keras.preprocessing.image.load_img(image_path, target_size=(224, 224))
    img_data = tf.keras.preprocessing.image.img_to_array(img)
    img_data = np.expand_dims(img_data, axis=0)
    img_data = tf.keras.applications.vgg16.preprocess_input(img_data)
    features = model.predict(img_data)
    features = features.flatten()
    return features

def recommend_similar_images(query_image_path, image_paths, top_k=5):
    query_features = extract_features(query_image_path)
    all_features = [extract_features(img_path) for img_path in image_paths]
    similarities = cosine_similarity([query_features], all_features)[0]
    top_indices = np.argsort(similarities)[-top_k:]
    return [image_paths[i] for i in top_indices]

# 示例使用
query_image = 'path_to_query_image.jpg'
image_database = ['path_to_image1.jpg', 'path_to_image2.jpg', ...]
recommended_images = recommend_similar_images(query_image, image_database)
print("Recommended images:", recommended_images)