no code implementations • 17 Mar 2023 • Alexander Hepburn, Valero Laparra, Raúl Santos-Rodriguez, Jesús Malo
Moreover, the direct evaluation of the hypothesis was limited by the inability of the classical image models to deliver accurate estimates of the probability.
no code implementations • 15 Mar 2023 • Jorge Vila-Tomás, Pablo Hernández-Cámara, Jesús Malo
In fact, the specific cancellation lights (and not the network architecture) are key to obtain human-like curves: results show that the classical choice of the lights is the one that leads to the best (more human-like) result, and any other choices lead to progressively different spectral sensitivities.
no code implementations • 26 Feb 2023 • Pablo Hernández-Cámara, Jorge Vila-Tomás, Valero Laparra, Jesús Malo
In this work, we perform a thorough analysis of the perceptual properties of pre-trained nets (particularly their ability to predict image quality) by isolating different factors: the goal (the function), the data (learning environment), the architecture, and the readout: selected layer(s), fine-tuning of channel relevance, and use of statistical descriptors as opposed to plain readout of responses.
no code implementations • 8 Jun 2022 • Valero Laparra, Alexander Hepburn, J. Emmanuel Johnson, Jesús Malo
Here we present the \emph{Convolutional RBIG}: an extension that alleviates this issue by imposing that the rotation in RBIG is a convolution.
1 code implementation • 25 Mar 2022 • Pablo Hernández-Cámara, Valero Laparra, Jesús Malo
In addition to the results on the Cityscapes and Foggy Cityscapes datasets, we explain these advantages through visualization of the responses: the equalization induced by the divisive normalization leads to more invariant features to local changes in contrast and illumination.
no code implementations • ICLR 2022 • Alexander Hepburn, Valero Laparra, Raul Santos-Rodriguez, Johannes Ballé, Jesús Malo
Since machine learning relies on the statistics of training data as well, the above connection has interesting implications when using perceptual distances (which mimic the behavior of the human visual system) as a loss function.
2 code implementations • 2 Dec 2020 • J. Emmanuel Johnson, Valero Laparra, Gustau Camps-Valls, Raul Santos-Rodríguez, Jesús Malo
Density destructors are differentiable and invertible transforms that map multivariate PDFs of arbitrary structure (low entropy) into non-structured PDFs (maximum entropy).
no code implementations • 21 Nov 2019 • Alexander Gomez-Villa, Adrian Martín, Javier Vazquez-Corral, Jesús Malo, Marcelo Bertalmío
Visual illusions are a very useful tool for vision scientists, because they allow them to better probe the limits, thresholds and errors of the visual system.
no code implementations • 28 Oct 2019 • Alexander Hepburn, Valero Laparra, Jesús Malo, Ryan McConville, Raul Santos-Rodriguez
Traditionally, the vision community has devised algorithms to estimate the distance between an original image and images that have been subject to perturbations.
no code implementations • 31 Jan 2016 • Valero Laparra, Juan Gutiérrez, Gustavo Camps-Valls, Jesús Malo
In this paper, we propose an alternative non-explicit way to take into account the relations among natural image wavelet coefficients for denoising: we use support vector regression (SVR) in the wavelet domain to enforce these relations in the estimated signal.
no code implementations • 31 Jan 2016 • Valero Laparra, Sandra Jiménez, Devis Tuia, Gustau Camps-Valls, Jesús Malo
Moreover, PPA shows a number of interesting analytical properties.
no code implementations • 31 Jan 2016 • Valero Laparra, Sandra Jiménez, Gustavo Camps-Valls, Jesús Malo
Here we address the simultaneous statistical explanation of (i) the nonlinear behavior of achromatic and chromatic mechanisms in a fixed adaptation state, and (ii) the change of such behavior.
7 code implementations • IEEE Transactions on Neural Networks 2011 • Valero Laparra, Gustavo Camps-Valls, Jesús Malo
The practical performance of RBIG is successfully illustrated in a number of multidimensional problems such as image synthesis, classification, denoising, and multi-information estimation.
no code implementations • 18 Oct 2013 • Gustavo Camps-Valls, Juan Gutiérrez, Gabriel Gómez-Pérez, Jesús Malo
We analytically demonstrate that no linear domain fulfills this condition because of the statistical and perceptual inter-coefficient relations that exist in these domains.