作者: Robert Templeman , Mohammed Korayem , David Crandall , Apu Kapadia
关键词:
摘要: As a first step, we introduce PlaceAvoider, technique for owners of first-person cameras to ‘blacklist’ sensitive spaces (like bathrooms and bedrooms). PlaceAvoider recognizes images captured in these flags them review before the are made available applications. performs novel image analysis using both fine-grained features specific objects) coarse-grained, scene-level colors textures) classify where photo was taken. combines probabilistic framework that jointly labels streams order improve accuracy. We test on five realistic firstperson datasets show it is robust blurriness, motion, occlusion.