Project Description:
The aim of our work project was the creation of automatic tool capable to detect and remove shadows from a single color image thereby obtaining a shadow-free image.
The shadow removal process consists of two challenging subtasks: detecting the shadowed region and restoring the illumination in that region.
We first identify shadowed and lit areas on the same surface in the scene using an illumination-invariant distance measure.
Then we automatically compute a mask that indicates the lit part of the surface and another mask that indicates the entire shadowed area.
The next step is the illumination intensity recovery at the shadowed pixels, to do so we need to estimate the four parameters
of the affine model. In order to estimate these parameters we use two strips of pixels: one inside the shadowed region, and the other outside the region.
The last problem that we face is inpainting the borders of the shadow.
Hence, the problem area is typically reduced to a very thin strip of pixels along the boundary of the shadow mask which separate the shadowed and illuminated area.
|
