CPC G06T 13/205 (2013.01) [G06T 7/246 (2017.01); G06T 13/40 (2013.01); G06T 13/80 (2013.01)] | 20 Claims |
1. A method comprising:
identifying, using one or more processors of a user device, a plurality of animation elements within a computer model; wherein the computer model and the plurality of animation elements are generated by processing an image using an overlay template to generate the computer model and the plurality of animation elements;
for each animation element of the plurality of animation elements:
identifying a plurality of motion patterns; and
identifying a plurality of speed harmonics;
generating, using the one or more processors, a set of motion data values comprising a state-space description of the plurality of motion patterns and the plurality of speed harmonics for the plurality of animation elements within the computer model;
assigning, by the one or more processors, a probability to each value of the set of motion data values for the state-space description;
processing audio inputs to identify a set of audio characteristics for audio data received at a microphone of the user device;
wherein processing the audio inputs further comprises:
generating a set of normalized audio energy values based on a set of standard audio input characteristics;
selecting a minimum audio energy threshold and a maximum audio energy threshold;
calculating a model motion energy for each value of the set of motion data values for the state-space description using a first motion model, wherein the model motion energy is determined by adding a product of a moment of inertial of each child of each of the plurality of animation elements and an angular velocity of the animation elements;
matching the first motion model to the set of normalized audio energy values; and
assigning a zero probability to a set of values of the set of motion data values for the state-space description where a corresponding model motion energy is less than the minimum audio energy threshold or greater than the maximum audio energy threshold as matched from the first motion model to the set of normalized audio energy values.
|