I read that with the current implementation there are issues when geometric transformations are involved in the translation (e.g. dog to cat). Are there any ideas or future developments to overcome this issue and allow to translate images not only by color and texture, but also geometry?
I read that with the current implementation there are issues when geometric transformations are involved in the translation (e.g. dog to cat). Are there any ideas or future developments to overcome this issue and allow to translate images not only by color and texture, but also geometry?