作者: Noriji Kato , Masahiro Maeda , Hitoshi Ikeda
DOI:
关键词: Geography 、 Section (archaeology) 、 Artificial intelligence 、 Pixel 、 Position (vector) 、 Action (philosophy) 、 Object (computer science) 、 Relation (database) 、 Face (geometry) 、 Computer vision 、 Image (mathematics)
摘要: An action recognition apparatus acquires image data and outputs information representing an content of a human captured in the data. The includes face-portion specifying section, acquisition output section. section specifies first region including face portion from skin-color peculiar to pixel values region. at least one second predetermined body except for based on information. position relation between specified