Neural Photometry-Guided Visual Attribute Transfer.

We present a deep learning-based method for propagating spatially-varying visual material attributes (e.g., texture maps or image stylizations) to larger samples of the same or similar materials. For training, we leverage images of the material taken under multiple illuminations and a dedicated data augmentation policy, making the transfer robust to novel illumination conditions and affine deformations. Our model relies on a supervised image-to-image translation framework and is agnostic to the transferred domain; we showcase a semantic segmentation, a normal map, and a stylization. Following an image analogies approach, the method only requires the training data to contain the same visual structures as the input guidance. Our approach works at interactive rates, making it suitable for material edit applications. We thoroughly evaluate our learning methodology in a controlled setup providing quantitative measures of performance. Last, we demonstrate that training the model on a single material is enough to generalize to materials of the same type without the need for massive datasets.

SeamlessGAN: Self-Supervised Synthesis of Tileable Texture Maps.

Real-time graphics applications require high-quality textured materials to convey realism in virtual environments. Generating these textures is challenging as they need to be visually realistic, seamlessly tileable, and have a small impact on the memory consumption of the application. For this reason, they are often created manually by skilled artists. In this work, we present SeamlessGAN, a method capable of automatically generating tileable texture maps from a single input exemplar. In contrast to most existing methods, focused solely on solving the synthesis problem, our work tackles both problems, synthesis and tileability, simultaneously. Our key idea is to realize that tiling a latent space within a generative network trained using adversarial expansion techniques produces outputs with continuity at the seam intersection that can then be turned into tileable images by cropping the central area. Since not every value of the latent space is valid to produce high-quality outputs, we leverage the discriminator as a perceptual error metric capable of identifying artifact-free textures during a sampling process. Further, in contrast to previous work on deep texture synthesis, our model is designed and optimized to work with multi-layered texture representations, enabling textures composed of multiple maps such as albedo, normals, etc. We extensively test our design choices for the network architecture, loss function, and sampling parameters. We show qualitatively and quantitatively that our approach outperforms previous methods and works for textures of different types.

Color Control Functions for Multiprimary Displays I: Robustness Analysis and Optimization Formulations.

Color management for a multiprimary display requires, as a fundamental step, the determination of a color control function (CCF) that specifies control values for reproducing each color in the display's gamut. Multiprimary displays offer alternative choices of control values for reproducing a color in the interior of the gamut and accordingly alternative choices of CCFs. Under ideal conditions, alternative CCFs render colors identically. However, deviations in the spectral distributions of the primaries and the diversity of cone sensitivities among observers impact alternative CCFs differently, and, in particular, make some CCFs prone to artifacts in rendered images. We develop a framework for analyzing robustness of CCFs for multiprimary displays against primary and observer variations, incorporating a common model of human color perception. Using the framework, we propose analytical and numerical approaches for determining robust CCFs. First, via analytical development, we: (a) demonstrate that linearity of the CCF in tristimulus space endows it with resilience to variations, particularly, linearity can ensure invariance of the gray axis, (b) construct an axially linear CCF that is defined by the property of linearity over constant chromaticity loci, and (c) obtain an analytical form for the axially linear CCF that demonstrates it is continuous but suffers from the limitation that it does not have continuous derivatives. Second, to overcome the limitation of the axially linear CCF, we motivate and develop two variational objective functions for optimization of multiprimary CCFs, the first aims to preserve color transitions in the presence of primary/observer variations and the second combines this objective with desirable invariance along the gray axis, by incorporating the axially linear CCF. A companion Part II paper, presents an algorithmic approach for numerically computing optimal CCFs for the two alternative variational objective functions proposed here and presents results comparing alternative CCFs for several different 4,5, and 6 primary designs.

Color Control Functions for Multiprimary Displays II: Variational Robustness Optimization.

In a companion Part I paper, we presented a framework for analyzing robustness of color control functions (CCFs) for multiprimary displays against primary and observer variations and proposed a variational minimization for obtaining robust CCFs. The objective function proposed in the Part I paper combines two nonnegative terms that serve as useful figures of merit for quantitatively characterizing CCFs. The first term measures lack of smoothness of the CCFs and characterizes how well transitions in perceptual color space are preserved in the presence of the primary/observer variations. The second term measures deviation of the CCF, in the vicinity of the gray axis, from a specific axially linear CCF that provides perceptual invariance to the variations along the gray axis. In this paper, using calculus of variations, we develop an algorithm for numerically computing optimal CCFs under the proposed variational formulation. Using the proposed algorithm, we determine optimal CCFs for a several multiprimary display designs and assess and compare their performance against alternative approaches. The variationally optimal CCFs obtained using the proposed approach offer improvements over the alternatives, as assessed visually and via quantitative metrics measuring smoothness and invariance in the presence of primary variations. The relative improvements provided by the proposed CCF increase with increasing number of primaries.

Decision tree learning to predict overweight/obesity based on body mass index and gene polymporphisms.

The new technologies for data analysis, such as decision tree learning, may help to predict the risk of developing diseases. The aim of the present work was to develop a pilot decision tree learning to predict overweight/obesity based on the combination of six single nucleotide polymorphisms (SNP) located in feeding-associated genes. Genotype study was performed in 151 healthy individuals, who were anonymized and randomly selected from the TALAVERA study. The decision tree analysis was performed using the R package rpart. The learning process was stopped when 15 or less observation was found in a node. The participant group consisted of 78 men and 73 women, who 100 individuals showed body mass index (BMI) ≥ 25 kg/m2 and 51 BMI < 25 kg/m2. Chi-square analysis revealed that individuals with BMI ≥ 25 kg/m2 showed higher frequency of the allelic variation Ala67Ala in AgRP rs5030980 with respect to those with BMI <25 kg/m2. However, the variant Thr67Ala in AgRP rs5030980 was the most frequently found in individuals with BMI <25 kg/m2. There were no statistical differences in the other analyzed SNPs. Decision tree learning revealed that carriers of the allelic variants AgRP (rs5030980) Ala67Ala, ADRB2 (rs1042714) Gln27Glu or Glu27Glu, INSIG2 (rs7566605) 73 + 9802 with CC or GG genotypes and PPARG (rs1801282) with the allelic variants of Ala12Ala or Pro12Pro, will most likely develop overweight/obesity (BMI ≥ 25 kg/m2). Moreover, the decision tree learning indicated that age and gender may change the developed three decision learning associated with overweight/obesity development. The present work should be considered as a pilot demonstrative study to reinforce the broad field of application of new data analysis technologies, such as decision tree learning, as useful tools for diseases prediction. This technology may achieve a potential applicability in the design of early strategies to prevent overweight/obesity.