A new colour 3D SFS scheme using neural networks and iterative recursive method

The problem of colour SFS (shape from shading) refers to the well-known fact that most real objects usually contain mixtures of diffuse and specular colour reflections and are affected by the multi-coloured interreflection under unknown reflectivity. In this paper, a novel colour SFS scheme is proposed. The proposed approach focuses on two main parts. First, a generalized neural based colour reflectance model is described. Second, an iterative recursive method is developed to reconstruct a multi-colour 3D surface. Experimental results on synthetic coloured objects were performed and compared with other well-known conventional models. The obtained results demonstrate that the performance of the proposed methodology is very encouraging.