Deep Photo Rally: Let's Gather Conversational Pictures

Author: Kazuki Ookawara, Hayaki Kawata, Masahumi Muta, Soh Masuko, Takehito Utsuro, Jun'ichi Hoshino


In this paper, we propose an anthropomorphic approach to generate speech sentences of a specific object according to surrounding circumstances using the recent Deep Neural Networks technology. In the proposal approach, the user can have pseudo communication with the object by photographing the object with a mobile terminal. We introduce some examples of application of the proposal approach to entertainment products, and show that this is an anthropo‐morphic approach capable of interacting with the environment

Copied! instagram
Research Areas : #Vision Program
Careers : Open Positions