A secure retrieval method suitable for large-scale images in cloud environment
A cloud environment, large-scale technology, applied in computer security devices, digital data information retrieval, character and pattern recognition, etc., can solve the problems of security, accuracy and efficiency can not be balanced, to improve retrieval efficiency and increase security , to achieve the effect of safe retrieval
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Publication Date
- 2021-05-04
Smart Images

Figure 1 
Figure 2 
Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the field of multimedia information security protection, and in particular relates to an image security retrieval method based on a bag-of-words model combined with a minimum hash principle, which can be used for security retrieval of large-scale images. Background technique
[0002] With the popularization of digital cameras and smart phones, people's access to data has become more and more convenient, and multimedia data such as images has shown an explosive growth trend. A cloud computing platform that integrates grid, parallel processing, and distributed processing provides a strong guarantee for massive data services and application processing with its low cost, powerful computing capabilities, and nearly unlimited resource pools. More and more of users choose to upload image data to a cloud server for storage and processing. However, data outsourced to the cloud is completely out of the direct physical control of its own...
Examples
Embodiment Construction
[0064] The present invention provides an image security retrieval method based on the bag-of-words model in a cloud environment. The specific implementation steps are as follows:
[0065] Step 1. Establish a bag-of-words model based on the image database to generate a visual dictionary and a median matrix. Specifically include the following sub-steps:
[0066] Step 1.1, local feature extraction: for each image in the image library, use the sift feature extraction algorithm to extract image features and generate feature point descriptors;
[0067] Step 1.2, build a visual dictionary: use the k-means clustering algorithm to train the feature points in the image training data set to generate k cluster centers, and each cluster center is represented as a visual word, which constitutes a k-dimensional visual dictionary W ;
[0068] Step 1.3, construct the median matrix: calculate the median value in each dimension for the image feature vectors belonging to the c(c∈[1,k])th cluste...