I think you needed more time and a more sophisticated approach. One could have used a 3D scanner and acquired the body shape and then one could parameterize it and then figure out the deformation you want to achieve from the bra -- even preview it for the customer, where do they want things to end up, and how much stress that will put on their body -- and then one could figure out the patterns to make again from the contact areas you wanted.
Basically it is a 3D problem and I think you didn't take a sophisticated approach because you are likely not knowledgable in that area.
You tried to apply tailoring to it, but tailoring is not body fitting nor shape changing as a bra is, nor is tailoring about stresses and weight distribution (which is comfort.)
It is a solvable problem, but your approach was not the correct one, it was simplistic.
Further, by utilizing generated (synthetic) data with 3d scanned human models, or generating your own with something like makehuman or daz3D, you should be able to find a way to UV unwrap a distorted sphere. I've had a lot of success with generating synthetic data like this with Blender 3D and training various neural networks (Mask RCNN / Deeplab for example) with various combinations of this synthetic and real datasets.
Comments
I think you needed more time and a more sophisticated approach. One could have used a 3D scanner and acquired the body shape and then one could parameterize it and then figure out the deformation you want to achieve from the bra -- even preview it for the customer, where do they want things to end up, and how much stress that will put on their body -- and then one could figure out the patterns to make again from the contact areas you wanted.
Basically it is a 3D problem and I think you didn't take a sophisticated approach because you are likely not knowledgable in that area.
You tried to apply tailoring to it, but tailoring is not body fitting nor shape changing as a bra is, nor is tailoring about stresses and weight distribution (which is comfort.)
It is a solvable problem, but your approach was not the correct one, it was simplistic.
It is an incredibly interesting problem though.
I agree with your take on this. Facebook has done considerable work with their Detectron / Dense Pose projects which can UV unwrap human bodies using AI. (https://github.com/facebookresearch/DensePose) Another is the smpl / smplify-x project (https://github.com/vchoutas/smplify-x)
Further, by utilizing generated (synthetic) data with 3d scanned human models, or generating your own with something like makehuman or daz3D, you should be able to find a way to UV unwrap a distorted sphere. I've had a lot of success with generating synthetic data like this with Blender 3D and training various neural networks (Mask RCNN / Deeplab for example) with various combinations of this synthetic and real datasets.
Yup, they have the wrong tech approach. Good they stopped wasting money though…