This paper describes the design and evaluation of AffectCam, a wearable system integrating SenseCam and BodyMedia SenseWear for capturing galvanic skin response as a measure of bodily arousal. AffectCam’s algorithms use arousal as a filtering mechanism for selecting the most personally relevant photos captured during people’s ordinary daily life, i.e. high arousal photos. We discuss initial findings showing that emotional arousal does improve the quality of memory recall associated with emotionally arousing events. In particular, the high arousal photos support richer recall of episodic memories than low arousal ones, i.e. over 50% improvement. We also consider how various memory characteristics such as event itself together with emotions and thoughts at the time of encoding, as well as its spatio-temporal context are differently cued by the AffectCam.