Wednesday 2/13 CSB 632 10:00 AM The OD Theory of TOD: The Use and Limits of Temporal Information for Object Discovery Speaker: Brandon Sanders Abstract: We present the theory behind TOD (the Temporal Object Discoverer), a novel unsupervised system that uses only temporal information to discover objects across image sequences acquired by any number of uncalibrated cameras. The process is divided into two stages: (1) Construction of a global event schedule; (2) Labeling of observations using the global schedule. Our Global Scheduling (GSched) algorithm constructs a global schedule when certain observability criteria are met. Our Quasi-Static Labeling (QSL) algorithm uses the schedule created by GSched to produce the maximally-informative mapping of each pixel's observations onto the objects they stem from. Using GSched and QSL, TOD ignores distracting motion, correctly deals with complicated occlusions, and naturally groups observations across cameras. The sets of 2D masks TOD recovers are suitable for unsupervised training and initialization of object recognition and tracking systems.