What are the iris recognition security issues?

The demand for secure and trusted identity authentication is growing rapidly, but the emphasis on the security of biometric images and feature templates is far from enough. Biometric features, including iris features, have life-changing characteristics, which indicates that once lost, it means that it is lost, there is no way to re-edit or publish a template to make up for the loss. The existence of these security threats is seriously restricting the further development based on biometrics.

At present, during the application of iris, the safety hazard mainly manifests as the leakage of the original iris image and the image restoration of the iris feature template. The details are as follows:

(1) Leakage of the original iris image: The collection, transmission, storage, comparison and other aspects of the iris application process in the gong area, the original image of the iris is a concentrated expression of personal privacy data, and there is a risk of privacy leakage. When a large area of ​​leakage occurs, it will cause serious social impact.

(2) Image restoration of iris feature template: iris feature template is restored to iris image for other applications by some algorithms. Although the restored image is not exactly the same as the original image, it can also be used for iris recognition or other iris algorithm. .

For the security risks of the above iris application, it is necessary to study the original image and feature template protection scheme based on cryptography to protect the iris data security to meet the security requirements of users in the process of identity collection, storage and authentication. The lifetime invariance of the iris characteristics and the massive data in the system iris image feature database pose great challenges to the research of the technology, especially in the critical requirements of security technologies such as correct matching rate, false rejection rate, and authentication speed. Next, it is a severe test of whether the technology can achieve further development.

For biometric security, domestic and foreign research institutions are mainly divided into two major technical genres: data encryption and data obfuscation: the method based on data encryption is to bind the key and image/feature information, generate ciphertext for transmission, and play private data. The purpose of protection, but it involves the issue of symmetric key distribution. Data confusion is mainly based on template deformation technology, random number session multi-factor obfuscation technology, digital watermarking, biometric data mosaic, etc. See Figure 3 for the basic method of iris data protection. Both methods cause some loss to the efficiency of iris acquisition, transmission, and comparison.