I have several datasets with faces, which are manually categorized by human experts based on their facial features, such as nasolabial fold, forehead wrinkles, under eye bags, neck, area between eyebrows, lips, etc. I.e. mostly facial imperfections. For example, forehead wrinkles could be classified as categories from 1 to 5 by severity. There are 10 face areas that we target. We specialize in face gymnastics, i.e. prescribe exercises to improve a particular feature. For example, if we categorize your forehead wrinkles as category N=3, we can correct that with exercises to category N=2. We would like to be able to morph a face from its current category N into a category N - 1. I suppose it won't make much more effort to morph into any category number. So, the project is to programmatically morph a face from its current category given input from a different category. Again, this applies to 10 areas.
The input: a selfie photo, area of the face to morph, current category
The output: the selfie photo with the category - 1 of the face area
I have ~800 total selfie-style photos divided into 5 categories for each face area. Some of the categories are unbalanced. For example, category 5 is the most wrinkled, so we have the smallest amount of images.
Could you provide a quote for this project? Thank you.
9 freelancers are bidding on average $949 for this job
Hello, I have checked The requirement and understand that as well. I have created SIMILAR in past. Do you want to see the DEMO ??? Can will send you ?? Thanks.