I wonder why there is so little research on creating/analyzing vector art and vectorising existing images using ANNs. There are so much possibilities there: us

I wonder why there is so little research on creating/analyzing vector art and vectorising existing images using ANNs. There are so much possibilities there: using closed circuits or gestalt blobs for colorizing, flawless style transfer and art/non-art separation, generating detailed sketches from words and rough drawings etc.