AUTOMATICALLY DETECTING "SIGNIFICANT EVENTS" ON SenseCam

摘要: SenseCam is an effective memory-aid device that can automatically record images and other data from the wearer's whole day. The main issue that, while produces a sizeable collection of over time period, vast quantity captured contains large percentage routine events, which are little interest to review. In this article, aim detect "Significant Events" for wearers. We use several series analysis methods such as Detrended Fluctuation Analysis (DFA), Eigenvalue dynamics Wavelet Correlations analyse multiple generated by SenseCam. show exposes strong long-range correlation relationship in collections. Maximum Overlap Discrete Transform (MODWT) was used calculate equal-time Correlation Matrices different scales then explore granularity largest eigenvalue changes ratio sub-dominant spectrum sliding windows. By examination eigenspectrum, we these approaches enable detection major events recording, with MODWT also providing useful insight on details events. suggest some wavelet (e.g., 8 minutes–16 minutes) have potential identify distinct or activities.