Visualizing Personal Lifelog Data for Deeper Insights @ NTCIR-13 Lifelog-2 Cognitive Vision Lab, Department of Visual Computing Institute for Infocomm Research (I 2 R), A*STAR We aim to generate and visualize lifelog in- sights under the NTCIR -13 Lifelog-2, Lifelog In- sight (sub) Task (LIT). Authors: Q.L. Xu ([email protected]), V. Subbaraju, A.G. del Molino, J. Lin, F. Fang, J-H Lim, L. Li, V. Chandrasekhar Features (Deep Learning) Activity Annotation Training (Retrieval) Cluster Aggregate Correlate Animate Logging Data Compare External Data Retrieval Insight Mining Ground truth by Annotation Visualize Browse Search Narrate Advise • Provide insights into the diet and blood sugar levels of the lifeloggers. Diet • Describe the exercise, sleep and physical activities of both lifeloggers Exercise • Socialisation levels are a good indicator of the health of individuals Social • Provide insights onto the location and movement patterns of the lifeloggers Where • Comparison between two individuals across multiple dimensions Compare Aim Method Visualization Templates Mobile App User Interface Generate minute-wise annotation of the users’ activities. Generate insights of users’ activities according to a suite of templates. Build a prototype mobile app to vis- ualize the insights.