An important advice concerning the use of AI generated objects from pictures: the mesh's model is dense, some nice details are lost and the textures extracted from the picture have a poor quality. (I don't take into account that the generation of hidden parts might generate additional limbs, fingers, etc.). That's why it's better to use them as far static objects into the scene, partly hidden with a DOF.

But I didn't follow this advice!

I have added the two pictures used to generate the static characters of this render.