Tsinghua University | 中文版

Research News

Dynamic security code speech based unsupervised identity authentication system of RIIT acquired the international leading evaluation of scientific research achievements of China Electronic Society

On May 23, the “dynamic security code speech based unsupervised identity authentication system " scientific and technological achievements Appraisal Conference, organized by the Chinese Electronic Society, was held in RIIT. The appraisal conference was presided by Professor Gao Wen, academician of the Chinese Academy of Engineering, Director of the Department of Information and Engineering Science of Peking University, Chairman of the Chinese Computer Society, and former Deputy Director of the National Natural Science Foundation of China. Cao Yuhong, director of the Science and Technology Evaluation and Achievement Transformation Center of the Chinese Electronic Society, and Ji Yindong, president of RIIT, attended the meeting.



The project team made a detailed research report, technical report and video demonstration to the appraisal committee, and reported the latest results and its application in the fields of finance, social security, automobile, etc. Finally, they made the economic and social benefits analysis from the direct and indirect, and the expected economic benefits. Subsequently, the test team announced the pre-test opinions and the name list of test groups, and the project team reported the preliminary testing work, then accepted that the results of the acceptance test were basically consistent with the test results in the report--- the system has strong noisy robustness and self-learning ability.

After questioning and discussion by the expert group, the appraisal committee unanimously decided that the result satisfied the demand of technologies and applications of identity authentication based on voiceprint in the case of being unsupervised and reached the international leading level. In view of the security in protecting users’ identity authentication of the results with low cost, high accuracy and weak privacy, the Committee recommends that the application be promoted in all fields as soon as possible.

The “dynamic security code speech based unsupervised identity authentication system " was jointly achieved by Speech and Language Technology R&D Center and Beijing d-Ear Technologies Co., Ltd.. Under unsupervised circumstance in cyberspace, the achievement utilized the speech signal’s characteristic of ‘simple in form, rich in information’, integrated voiceprint recognition, speech recognition, emotion recognition, anti-recording detection, etc., presented dynamic security code speech based unsupervised identity authentication technology scheme, and accurately implemented unsupervised identity authentication, providing new technologies and applications of Internet, networking and big data as well as security protection for people’s daily life. The products using the key technologies of this achievement have been applied in the fields of finance, social security, medical insurance and security, which have brought about great economic and social benefits, and are typical cases of artificial intelligence application. It is of great significance for the construction of smart city and the development of social economy.