Data Fusion in Shape From Shading Algorithms
The vision process that we examine in this chapter is that of obtaining object shape from information about the specular and Lambertian components of the light reflected off of the object. We demonstrate the application of two different forms of weakly coupled fusion to the problem of estimating object shape from shading information. The first example is of an algebraic approach, while the second involves an energy function or Bayesian approach. We follow this description of the weakly coupled methods with an illustration of how we can convert the weakly coupled fusion algorithms into a strongly coupled fusion algorithm.
KeywordsMinimal Surface Data Fusion Image Component Smoothness Constraint Smoothness Term
Unable to display preview. Download preview PDF.