The present research was inspired on the seminal work
by Scott Johnston [8] that uses image-based techniques. He
also estimates normal vectors and uses sparse interpolation
to approximate them over the remaining image to illuminate
2D drawings. However, the present work has major conceptual and technical differences. The main drawbacks of
the Johnston’s method are the need of multi-channel images
and the need of scene transformation to vectorial space. In
our model, we work directly from the 2D image and we
avoid transformations to vectorial space completely. This is
the reason why we claim that our method is the only one in
the literature that is genuinely an image-based method.