In preparation of subsequent blogs about my research in machine learning.
An overview of the project.
vertÅ (â€œâ€˜turnâ€™â€) + pellis ( â€œhideâ€) is Latin for Shapeshifter / Werewolf.
As its name denotes, it is a system that can mimic a sentient being starting with its voice and then visual features. It learns and assimilates from all types of relevant content. This can be but is not limited to phonic and visional content.
The grand vision is that it can be fed images, videos, and audio files of a subject and it would learn to mimic said subject through voice, visual etc.
Create an online platform that can speak the text input. The platform will learn from media starting with video files. The first goal is to learn from the audio from the video files. Then the second goal will be to learn the visual representation synced to the audio file.
Once the learning is completed, users can input text and hear spoken in the voicer VP learned from.
Long term goals is a platform that could mimick or shapeshift into any personality or person that the user points it to; given enough media content of course.