Rebooting my 3D reconstruction project

My undergrad thesis project was based on 3D reconstruction of an object from a single image. (Using a text prompt to classify the object)

Recently I have been thinking of rebooting the project as a web-app.

These are my thoughts:

Reconstruction can only be done for objects. (Or should I limit it to objects for now)

so let the user select the object -> with text prompt? through GUI?

Let the image or object be generated -> GenAI (Hasn’t been integrated yet)

FLOW :=> Image -> Object -> Mask -> Depth -> Point Cloud

Q. If I segment the parts of the object seperately then reconstruct their 3D models seperately and then put the reconstructions back together would that work better?

Q. From sales perspective should I work towards marketing the software output i.e. the models for AR, VR or 3D printer or all? basically like templates