Shrinkage fields is a random field-based machine learning technique that aims to perform high quality image restoration (denoising and deblurring) using low computational overhead.
The restored image
is predicted from a corrupted observation
after training on a set of sample images
A shrinkage (mapping) function
π
π
γ 2
μ
is directly modeled as a linear combination of radial basis function kernels, where
is the shared precision parameter,
μ
denotes the (equidistant) kernel positions, and M is the number of Gaussian kernels.
Because the shrinkage function is directly modeled, the optimization procedure is reduced to a single quadratic minimization per iteration, denoted as the prediction of a shrinkage field
λ
π
λ
denotes the discrete Fourier transform and
with point spread function filter,
is an optical transfer function defined as the discrete Fourier transform of
is the complex conjugate of
with the initial case
, this forms a cascade of Gaussian conditional random fields (or cascade of shrinkage fields (CSF)).
Loss-minimization is used to learn the model parameters
π
The learning objective function is defined as
is a differentiable loss function which is greedily minimized using training data
Preliminary tests by the author suggest that RTF5[1] obtains slightly better denoising performance than
BM3D denoising speed falls between that of
, RTF being an order of magnitude slower.