The application discloses a kind of individualized tone and image
robot, belong to
artificial intelligence and man-
machine interaction technical field.The
robot includes voice acquisition module, tone
feature extraction module, individualized tone
library,
image acquisition module, image
feature extraction module, individualized image
library, character
profile management unit, built-in audio-visual synchronization engine output control module, switching module and controller.The application stores tone model and image model by character
profile management unit Binding, realize one-key
synchronous switching;Through audio-visual synchronization engine, image
mouth shape, expression and action are driven in real
time based on voice timing, and the audio-visual synchronization error is controlled within 100 milliseconds;Combined with
small sample feature enhancement and the collaborative architecture of cloud training local
inference, the use threshold and hardware cost are reduced.The application can be widely applied to
child care,
home entertainment and other scenes, and significantly improves the realism, immersion and emotional temperature of human-computer interaction.