Methods and systems for advanced and augmented training of deep neural networks using synthetic data and innovative generative networks

A system for a server comprising:

• a processing core having a deep neural network (DNN);
• an I/O hub controller coupled to the processing core and to provide network, data storage, and DNN access for the processing core;
• and
• a graphics processor coupled to the I/O hub controller and to
• train a DNN using synthetic data,
• train a plurality of DNNs using context data,
• associate features of the DNNs trained using context data with features of the DNN trained with synthetic data, and
• generate an augmented DNN using the associated features.

A deep neural network (DNN) training method comprising:

• processing synthetic labeled images having labeled information to generate visually realistic images;
• processing real unlabeled images and the visually realistic images generated using labeled information from synthetic labeled images;
• and
• discriminating between the processed real unlabeled images and the visually realistic images to determine if the visually realistic images are real data or not real data.