Method for generating objects using an hourglass predictor

A computer-implemented method for transforming an object, comprising:

• training an autoencoder on training input data to identify latent variables and generate training output data, where the autoencoder comprises an encoder and a decoder, where the encoder converts the training input data into the latent variables, where the decoder converts the latent variables to training output data, where the training output data is at least substantially the same as the training input data, and where the trained autoencoder encodes input data into the latent variables with the encoder and decodes the latent variables into output data with the decoder;
• identifying a transformation function with at least one data pair, where each of the at least one data pair comprises an initial object in an initial state and a transformed object in a transformed state, by:
• generating initial state latent variables for each said initial object by using a representation of each said initial object as the input data for the encoder;
• generating transformed state latent variables for each said transformed object by using a representation of each said transformed object as input data for the encoder, and identifying the transformation function based on the differences between the initial state latent variables and the transformed state latent variables;
• and
• transforming a new object with the transformation function by:
• generating new latent variables by using a representation of the new object as the input data for the encoder;
• generating transformed new latent variables by applying the transformation function to the new latent variables;
• generating transformed output data by decoding the transformed new latent variables with the decoder;
• where the initial object is an original tooth and the transformed object is an aged version of the original tooth.

A computer-implemented method for transforming an object, comprising:

• training an autoencoder on training input data to identify latent variables and generate training output data, where the autoencoder comprises an encoder and a decoder, where the encoder converts the training input data into the latent variables, where the decoder converts the latent variables to training output data, where the training output data is at least substantially the same as the training input data, and where the trained autoencoder encodes input data into the latent variables with the encoder and decodes the latent variables into output data with the decoder;
• identifying a transformation function with at least one data pair, where each of the at least one data pair comprises an initial object in an initial state and a transformed object in a transformed state, by:
• generating initial state latent variables for each said initial object by using a representation of each said initial object as the input data for the encoder;
• generating transformed state latent variables for each said transformed object by using a representation of each said transformed object as input data for the encoder, and identifying the transformation function based on the differences between the initial state latent variables and the transformed state latent variables;
• and
• transforming a new object with the transformation function by:
• generating new latent variables by using a representation of the new object as the input data for the encoder;
• generating transformed new latent variables by applying the transformation function to the new latent variables;
• generating transformed output data by decoding the transformed new latent variables with the decoder;
• where the initial object is an original tooth and the transformed object is a tooth with increased angular symmetry compared to the original tooth.