Multi-source ciphertext image retrieval method based on federated learning and secret sharing

Abstract

The invention discloses a multi-source ciphertext image retrieval method based on federated learning and secret sharing. The method comprises the following steps that: S1, model training is performed on a convolutional neural network of a double-cloud platform by taking an image owner joining the double-cloud platform as a federated member based on federated learning; and S2, the authorized user completes ciphertext image retrieval by means of the double-cloud platform based on additive secret sharing. According to the scheme, a multi-source ciphertext image retrieval scheme based on federated learning and secret sharing is provided, a neural network model structure used by federated learning simplification retrieval is utilized to obtain better network parameters, better neural network parameters and a simpler network model structure are obtained through overhead of the image owner end, a better convolutional neural network can be used in ciphertext image retrieval, and a user can obtain a better result faster during retrieval.

Description

technical field [0001] The invention belongs to the technical field of ciphertext image retrieval, and in particular relates to a multi-source ciphertext image retrieval method based on federated learning and secret sharing. Background technique [0002] With the increasing popularity of electronic imaging equipment, image resources have increased significantly, people began to have the need to find similar images, and the Internet has also begun to provide similar services, such as Baidu and other services that provide image search services. However, high-quality image resources are of high value and should not be directly published on the Internet. Therefore, it is a practical task to consider a solution to support image owners to outsource their own high-value images to the cloud platform, and at the same time ensure that the cloud platform can support retrieval services for these ciphertext images without knowing the content of the images. The existing ciphertext image ...

AI Technical Summary

Problems solved by technology

[0006] (1) Does not support multi-source ciphertext image retrieval, all existing single-cloud platform solutions cannot fully meet the functional requirements of multi-source ciphertext image retrieval, and all have the problem of low accuracy

Although the multi-cloud platform solution is easy to expand into a multi-source solution, the existing strategies still generally have problems such as high time consumption or insufficient security

[0007] (2) The retrieval accuracy is insufficient. The existing optimal ciphertext image retrieval scheme considers the convolutional layer features of the pre-trained neural network, which still has a large room for improvement in accuracy under large samples, while the rest of the traditional features or statistics Feature schemes have problems with accuracy even with small samples

[0008] (3) Time consumption is high. Among the existing high-precision solutions based on dual-cloud platforms, the technology based on secret sharing strategy is the lowest consumption. high communication time

Images