Abstract
In this paper, a technique of photo retrieval / indexing using human brain activities is presented. The non-invasive electroencephalogram (EEG) recording was applied to have event-related P300 evoked potential during photo retrieval task in oddball paradigm. Three subjects tried to select the photo images among nine that they want to pickup according to their mental states. The features of the P300 evoked potential were successfully extracted from temporal average EEG signals. It was found that the average performances of target photo selections reached more than 80% using three-channel EEG activities with four-time average. This concept will be applicable in future to achieve intuitive retrieval and indexing for life log with large quantities of data.
