Lightness and retinex theory.

By E. H. Land, J. McCann. 1971

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

3D Shape Induction from 2D Views of Multiple Objects (2016, background): “Optimization-based approaches put priors on geometry, material, and light and estimate all of them by minimizing the reconstruction error when rendered [17, 3, 2].”
Learning Non-Lambertian Object Intrinsics Across ShapeNet Categories (2016, background): “Land and McCann [18] observe that large gradients in images usually correspond to changes in reﬂectance and small gradients to smooth shading variations.”
Recovering intrinsic images from a single image (2005, background): “An early algorithm which attempted to recover the reflectance and illumination of a scene is the Retinex algorithm [14].”
Color Constancy, Intrinsic Images, and Shape Estimation (2012, methodology): “This insight is fundamental to most intrinsic image algorithms [2, 4, 5, 14, 23], and is used in our previous works [1, 22].”
Intrinsic Decompositions for Image Editing (2017, unknown): “This states that it is unlikely for a single image edge to result from both a change of illumination and reﬂectance at the same time [LM71].”
Deriving intrinsic images from image sequences (2001, background): “Land and McCann’s Retinex algorithm [7] could successfully decompose scenes in which the reflectance image was piecewise constant.”
Person re-identification by Local Maximal Occurrence representation and metric learning (2014, unknown): “Local Maximal Occurrence Feature 3.”
Illumination compensation and normalization for robust face recognition using discrete cosine transform in logarithm domain (2006, background): “In a simple situation, the image gray level f(x; y) can be assumed to be proportional to the product of the reflectance r(x; y) and the illumination e(x; y) [23], i.e., f(x; y) = r(x; y) 1 e(x; y): (1) To our knowledge, the Retinex algorithm is related to the reflectance constancy [17].”
Multiscale Retinex (2014, methodology): “The original retinex algorithm, proposed by Land [13], considers all possible paths starting at random points and ending at the pixel at which the lightness value is computed.”
A new algorithm for unsupervised global and local color correction (2003, background): “Channel independence, Gray World and White Patch mechanisms do not account for all the aspects of visual appearance: we need to also take into account spatial relationships in the scene, like other models do (McCann, 1999; Land and McCann, 1971; Spitzer and Sherman, 2002).”
Compressing and companding high dynamic range images with subband architectures (2005, methodology): “An alternate approach is to work in the gradient domain, as is done in Retinex algorithms [Land and McCann 1971].”
Self-Supervised Intrinsic Image Decomposition (2017, methodology): “as a further extension of these models which aims to relax the need for a complete set of ground truth data by modeling the image combination process, as in Nalbach et al. [2017]. Incorporating a domain-specific decoder to reconstruct input images has been explored by Hinton et al.”
CGIntrinsics: Better Intrinsic Image Decomposition through Physically-Based Rendering (2018, background): “The classical approach to intrinsic images is to integrate various priors (smoothness, reﬂectance sparseness, etc.) into an optimization framework [13,14,15,16,17,5].”
Face Liveness Detection from a Single Image with Sparse Low Rank Bilinear Discriminative Model (2010, methodology): “Here we prefer a type of variational Retinex approach, in which the illuminance μ(x, y) is first principally sought within the total variational framework and the albedo ρ(x, y) is then estimated through Land’s Retinex formula [11].”