Fig. 2From: Lifelogging caption generation via fourth-person vision in a human–robot symbiotic environmentOur concept of lifelogging caption generation in a human–robot symbiotic environment. The system is composed of the first-person perspective from a human, the second-person perspective from a service robot, and the third-person perspective from an embedded camera. Fragmental visual concepts are adaptively weighted and processed to predicting each wordBack to article page