Not sure if I understand your question... Sorry.
If you import an ortho image Uv (one sprtie front) it will create a flat geometry... After this, if u upload a side image, the program will interpolate both images and find matching pixels to create the 3d model on volume... The third image (top) can define missing geometry constraints.
Alternatively you can upload directly a 6 sprites image (using any upload button) and it creates the whole geometry at one shot.
You can mix any images to see the results, try the online version and test it by yourself. (up on this page).