Disculpa, pero esta entrada está disponible sólo en Inglés Estadounidense. For the sake of viewer convenience, the content is shown below in the alternative language. You may click the link to switch the active language.

I’m taking a small break to document this (failed?) attempt at training a Recurrent Neural Network so that it would be able to interpolate a movement sequence between two given poses. For some reason the output is rotated and it doesn’t make much sense.

The human figure is given by my program, that uses a set of unit vectors for the directions of the limbs. These vectors are the output of the RNN…