Life Events Segmentation Based on Lifelog Recorded by Wearable Device

This paper is based wearable sensing system structure, the operational framework is designed on the mechanism of "Aware - Computing - Action". It means that we collect the data from wearable device (Aware), compute and analysis the data (Computing) and then feedback the information of separated life event to user for future use (Action). In the "aware" stage, we collect the data of "life log" from wearable device and take the picture from personal digital device (PDA) every 30 seconds. The data of "life log" is inserted into Exif (Exchangeable image file format) of the image taken from PDA and then upload to the cloud database by smart phone. In the "computing" stage, we divide the data into the separated life event according to multiple features. These features include not only the original data (GPS coordinates, data of motion, time, etc.) but also the computed information, for example: the speed computed from GPS coordinates and time, the separated life event analyzed from computing the different among the images, etc. Having the information of separated life event, we turn into the "action" stage. In the "action" stage, this cloud system makes contact with emergency contact person immediately in critical situation in order to take action. Otherwise, the separated life event is stored in the cloud database for query and other purposes.