1G-36 Memorable Moment Extraction using the Pupil Dilation Analysis

These days, with the launch of Google glass, life-log which is recording videos and photos everyday has become reality. However, the stored data without indexing is meaningless and unattractive for users. Moreover, there is a problem that it is impossible to ensure storage for large amounts of data. Therefore, efficient meaningful data selection and automatic indexing will be important to be applied in practical use. In order to solve this problem, we propose a method that analyzing pupillary size variation through a camera device mounted on customized eye-glass. We use of the fact that if there are emotional changes, the pupillary size change. Recognition of emotional change is quite effective to extract a meaningful visual information automatically from the continuous recording of the view. Proposed system measures degree of expansion of the pupil diameter and picks a moment of rapid size change of pupil with camera. There are numerous studies to determine the emotional changes based on physiological signal processing such as EEG, ECG or GSR. Previous research usually attach the sensors to the body of the subjects in order to obtain a biological signal data although that is cause of discomfort. Hence, we propose new method to measure the size of the pupil only using lightweight camera. This has benefit that pupillary response time is much shorter than other signal and it can be reduced the analysis time remarkably.