Multiscaled Cross-Correlation Dynamics on SenseCam Lifelogged Images

摘要: In this paper, we introduce and evaluate a novel approach, namely the use of cross correlation matrix Maximum Overlap Discrete Wavelet Transform (MODWT) to analyse SenseCam lifelog data streams. is device that can automatically record images other from wearer’s whole day. It significant challenge deconstruct sizeable collection into meaningful events for users. The cross-correlation was used, characterise dynamical changes in non-stationary multivariate images. MODWT then applied equal-time Correlation Matrices over different time scales used explore granularity largest Eigenvalue changes, ratio sub-dominant spectrum dynamics, sliding windows. By examination eigenspectrum, show these approaches identify “Distinct Significant Events” wearers. dynamics across multiple provide useful insight on details major logged