Research On Identification Computing Based On Event Evoked Potential

In the new era of information, very important personal identification. But the identification methods such as fingerprint, face, voice recognition, there are certain drawbacks. Electroencephalogram EEG features stimulates people to their own interests to produce and other conditions are different, the preliminary study suggests that, everyone see the EEG feature of EEG their photos and see other people when he characteristics of different photos. Therefore, through the study of EEG signals, design and application of photo visual stimulation method for identity recognition.The main work of the thesis includes;1, research on EEG and event-related potentials, and summarized using EEG exercise imaginable conclusion, as the identity recognition based on event evoked potentials do preparation.2, design of identity recognition experiment scheme based on event evoked potential, divided into3patterns to pretend, noise data.3, from attending the stranger photos and specific photos, observation noise, from the attention of others to observe the fake photos. Data superposition method to observe the ERP components differ among. Then the identity verification using various data matlab analysis method.4, according to each subject to design its classifier, from the subjects of EEG and the correct rate and false rate were analyzed, the correct classification ratio of90%on average, that identify better subjects. With the promotion of practical value.