A secure and efficient face recognition method based on depth learning and homomorphic encryption

A homomorphic encryption and deep learning technology, applied in homomorphic encryption communication, neural learning methods, character and pattern recognition, etc., can solve the problem that the SCiFI scheme is no longer safe, and achieve the effect of reducing data volume and simplifying calculations

Inactive Publication Date: 2019-01-04
中共中央办公厅电子科技学院
View PDF5 Cites 32 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Later, Luong et al. proposed a method of attacking the SCiFI scheme based on reconstructed fragmented faces, making the SCiFI scheme no longer safe

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A secure and efficient face recognition method based on depth learning and homomorphic encryption
  • A secure and efficient face recognition method based on depth learning and homomorphic encryption
  • A secure and efficient face recognition method based on depth learning and homomorphic encryption

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0046] The present invention will be described in further detail below in conjunction with the accompanying drawings and specific embodiments.

[0047] Overall process:

[0048] The client has the encrypted public key and the decrypted private key, and the index i that it wants to look up the data in the server. The server has an encrypted public key and a face recognition threshold t. A uniform threshold is set for faces in the entire database through experiments. exist figure 2 where i and j represent the indexes of 0 and 1 in the server-side vector v, respectively. For the extracted 128-bit feature vector, J is the number of 1 in the server-side vector v, which is less than or equal to 128, w i with w j is a vector value in the client, + h represents the multiplication operation, - h stands for subtraction, d H is the Hamming distance, r is a random number, X k is the corresponding 0,1 vector;

[0049] client:

[0050] After the client acquires the face image, it...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a safe and efficient face recognition method based on depth learning and homomorphic encryption. Each terminal is responsible for acquiring face images by using depth neural network, extracting face features from the acquired face images, extracting 128-bit feature data, encrypting the feature data by paillier encryption algorithm and sending the ciphertext data to the cloud. The server at a cloud end receives the ciphertext data from the terminal, and calculates the Hamming distance by the formula of efficient hidden Hamming distance in this ciphertext state, and finally returns the recognition result to the terminal by OT protocol, which effectively protects the privacy. In the process of facial feature data encryption and Hamming distance computation, a parallel computing mode is introduced, which can encrypt multiple facial feature data and compute the corresponding Hamming distance in parallel, thus greatly improving the efficiency of the whole scheme and achieving the goal of high efficiency, and the method is easy to be implemented by software.

Description

technical field [0001] The invention belongs to the fields of cryptography, computer vision, face recognition, and image blind operation, and specifically relates to a safe and efficient face recognition method based on deep learning (DNN) and homomorphic encryption. Background technique [0002] In recent years, with the gradual maturity of face recognition technology, the application of face recognition technology has become more and more extensive. With the emergence of cloud computing, the traditional face recognition scheme has been changed, and a large amount of face image data has begun to be stored in the cloud, and the face recognition program has also been deployed to the cloud. Obviously, this solution can support large-scale video surveillance applications, so it can be applied in many aspects of real life. The most typical application is suspect search. Special cameras can be installed in some important public places, such as train stations, bus stations, etc....

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06N3/08H04L9/00
CPCH04L9/008G06N3/084G06V40/168
Inventor 李晓东韩青金鑫
Owner 中共中央办公厅电子科技学院
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap