Minimum Entropy Light and Shading Approximation

Retinal fundus image constrast normalization using mixture of Gaussians

“…The multiplicative and additive biasing fields are denoted by S × and S + respectively. For fundus imagery, S + is assumed to be zero, although it can be estimated by other means [5], [7]. The biasing field is further assumed to be low-frequency and the true image to be more or less pice-wise constant.…”

Principles and Methods for Face Recognition and Face Modelling

“…If a fully 3D model is learnt for the recognition, such as the example 3D reconstructions shown in this chapter, then it is possible to use the extra information to deal better with poor or inconsistent illumination. See for example the authors' work on shading and lighting correction using entropy minimzation (Bhalerao, 2006).…”