Browsing by Author "Garces, Elena"

Now showing 1 - 8 of 8
  • No Thumbnail Available
    Item
    Convolutional Sparse Coding for Capturing High‐Speed Video Content
    (© 2017 The Eurographics Association and John Wiley & Sons Ltd., 2017) Serrano, Ana; Garces, Elena; Masia, Belen; Gutierrez, Diego; Chen, Min and Zhang, Hao (Richard)
    Video capture is limited by the trade‐off between spatial and temporal resolution: when capturing videos of high temporal resolution, the spatial resolution decreases due to bandwidth limitations in the capture system. Achieving both high spatial temporal resolution is only possible with highly specialized and very expensive hardware, and even then the same basic trade‐off remains. The recent introduction of compressive sensing and sparse reconstruction techniques allows for the capture of high‐speed video, by coding the temporal information in a single frame, and then reconstructing the full video sequence from this single‐coded image and a trained dictionary of image patches. In this paper, we first analyse this approach, and find insights that help improve the quality of the reconstructed videos. We then introduce a novel technique, based on (CSC), and show how it outperforms the state‐of‐the‐art, patch‐based approach in terms of flexibility and efficiency, due to the convolutional nature of its filter banks. The key idea for CSC high‐speed video acquisition is extending the basic formulation by imposing an additional constraint in the temporal dimension, which enforces sparsity of the first‐order derivatives over time.Video capture is limited by the trade‐off between spatial and temporal resolution: when capturing videos of high temporal resolution, the spatial resolution decreases due to bandwidth limitations in the capture system. Achieving both high spatial and temporal resolution is only possible with highly specialized and very expensive hardware, and even then the same basic trade‐off remains. .
  • No Thumbnail Available
    Item
    Exploring Appearance and Style in Heterogeneous Visual Content
    (2017-10-19) Garces, Elena
    There are multiple ways to capture and represent the visual world; a drawing, a photograph, or a video are a few examples of visual data that are very frequent nowadays. Despite the different nature of each domain, there is a common need to process and edit these data after its production for different purposes. For example, we might want to modify the materials and the illumination of an object in a photograph, or we might want to explore a huge collection of non labeled images. The solutions to these problems mainly depend on the amount of information we have as input: it is not the same to process a plain set of colored pixels, like a photograph, than a scene captured with a 3D laser scan and multiple cameras. Thus, the nature of the visual data will also determine the complexity of the model we can use for processing. In this thesis, we focus on creating alternative representations of the visual content which will facilitate posterior editing and exploration tasks. In particular, we will focus on conventional visual data like pictures, video sequences , and light fields; and we will explore two different aspects or these data, the appearance in real scenes and the style in artistic scenes.
  • Thumbnail Image
    Item
    Fully Convolutional Graph Neural Networks for Parametric Virtual Try-On
    (The Eurographics Association and John Wiley & Sons Ltd., 2020) Vidaurre, Raquel; Santesteban, Igor; Garces, Elena; Casas, Dan; Bender, Jan and Popa, Tiberiu
    We present a learning-based approach for virtual try-on applications based on a fully convolutional graph neural network. In contrast to existing data-driven models, which are trained for a specific garment or mesh topology, our fully convolutional model can cope with a large family of garments, represented as parametric predefined 2D panels with arbitrary mesh topology, including long dresses, shirts, and tight tops. Under the hood, our novel geometric deep learning approach learns to drape 3D garments by decoupling the three different sources of deformations that condition the fit of clothing: garment type, target body shape, and material. Specifically, we first learn a regressor that predicts the 3D drape of the input parametric garment when worn by a mean body shape. Then, after a mesh topology optimization step where we generate a sufficient level of detail for the input garment type, we further deform the mesh to reproduce deformations caused by the target body shape. Finally, we predict fine-scale details such as wrinkles that depend mostly on the garment material. We qualitatively and quantitatively demonstrate that our fully convolutional approach outperforms existing methods in terms of generalization capabilities and memory requirements, and therefore it opens the door to more general learning-based models for virtual try-on applications.
  • No Thumbnail Available
    Item
    Graph-Based Reflectance Segmentation
    (The Eurographics Association, 2021) Garces, Elena; Gutierrez, Diego; Lopez-Moreno, Jorge; Silva, F. and Gutierrez, D. and Rodríguez, J. and Figueiredo, M.
    Most of the unsupervised image segmentation algorithms use just RGB color information in order to establish the similarity criteria between pixels in the image. This leads in many cases to a wrong interpretation of the scene since these criteria do not consider the physical interactions which give raise to of those RGB values (illumination, geometry, albedo) nor our perception of the scene. In this paper, we propose a novel criterion for unsupervised image segmentation which not only relies on color features, but also takes into account an approximation of the materials reflectance. By using a perceptually uniform color space, we apply our criterion to one of the most relevant state of the art segmentation techniques, showing its suitability for segmenting images into small and coherent clusters of constant reflectance. Furthermore, due to the wide adoption of such algorithm, we provide for the first time in the literature an evaluation of this technique under several scenarios and different configurations of its parameters. Finally, in order to enhance both the accuracy of the segmentation and the inner coherence of the clusters, we apply a series of image processing filters to the input image (median, mean-shift, bilateral), analyzing their effects in the segmentation process. Our results can be transferred to any image segmentation algorithm.
  • Thumbnail Image
    Item
    How Will It Drape Like? Capturing Fabric Mechanics from Depth Images
    (The Eurographics Association and John Wiley & Sons Ltd., 2023) Rodriguez-Pardo, Carlos; Prieto-Martín, Melania; Casas, Dan; Garces, Elena; Myszkowski, Karol; Niessner, Matthias
    We propose a method to estimate the mechanical parameters of fabrics using a casual capture setup with a depth camera. Our approach enables to create mechanically-correct digital representations of real-world textile materials, which is a fundamental step for many interactive design and engineering applications. As opposed to existing capture methods, which typically require expensive setups, video sequences, or manual intervention, our solution can capture at scale, is agnostic to the optical appearance of the textile, and facilitates fabric arrangement by non-expert operators. To this end, we propose a sim-to-real strategy to train a learning-based framework that can take as input one or multiple images and outputs a full set of mechanical parameters. Thanks to carefully designed data augmentation and transfer learning protocols, our solution generalizes to real images despite being trained only on synthetic data, hence successfully closing the sim-to-real loop. Key in our work is to demonstrate that evaluating the regression accuracy based on the similarity at parameter space leads to an inaccurate distances that do not match the human perception. To overcome this, we propose a novel metric for fabric drape similarity that operates on the image domain instead on the parameter space, allowing us to evaluate our estimation within the context of a similarity rank. We show that out metric correlates with human judgments about the perception of drape similarity, and that our model predictions produce perceptually accurate results compared to the ground truth parameters.
  • No Thumbnail Available
    Item
    Intrinsic Light Field Images
    (© 2017 The Eurographics Association and John Wiley & Sons Ltd., 2017) Garces, Elena; Echevarria, Jose I.; Zhang, Wen; Wu, Hongzhi; Zhou, Kun; Gutierrez, Diego; Chen, Min and Zhang, Hao (Richard)
    We present a method to automatically decompose a light field into its intrinsic shading and albedo components. Contrary to previous work targeted to two‐dimensional (2D) single images and videos, a light field is a 4D structure that captures non‐integrated incoming radiance over a discrete angular domain. This higher dimensionality of the problem renders previous state‐of‐the‐art algorithms impractical either due to their cost of processing a single 2D slice, or their inability to enforce proper coherence in additional dimensions. We propose a new decomposition algorithm that jointly optimizes the whole light field data for proper angular coherence. For efficiency, we extend Retinex theory, working on the gradient domain, where new albedo and occlusion terms are introduced. Results show that our method provides 4D intrinsic decompositions difficult to achieve with previous state‐of‐the‐art algorithms. We further provide a comprehensive analysis and comparisons with existing intrinsic image/video decomposition methods on light field images.We present a method to automatically decompose a into its intrinsic shading and albedo components. Contrary to previous work targeted to two‐dimensional (2D) single images and videos, a light field is a 4D structure that captures non‐integrated incoming radiance over a discrete angular domain. This higher dimensionality of the problem renders previous state‐of‐the‐art algorithms impractical either due to their cost of processing a single 2D slice, or their inability to enforce proper coherence in additional dimensions. We propose a new decomposition algorithm that jointly optimizes the whole light field data for proper angular coherence.
  • Thumbnail Image
    Item
    NEnv: Neural Environment Maps for Global Illumination
    (The Eurographics Association and John Wiley & Sons Ltd., 2023) Rodriguez-Pardo, Carlos; Fabre, Javier; Garces, Elena; Lopez-Moreno, Jorge; Ritschel, Tobias; Weidlich, Andrea
    Environment maps are commonly used to represent and compute far-field illumination in virtual scenes. However, they are expensive to evaluate and sample from, limiting their applicability to real-time rendering. Previous methods have focused on compression through spherical-domain approximations, or on learning priors for natural, day-light illumination. These hinder both accuracy and generality, and do not provide the probability information required for importance-sampling Monte Carlo integration. We propose NEnv, a deep-learning fully-differentiable method, capable of compressing and learning to sample from a single environment map. NEnv is composed of two different neural networks: A normalizing flow, able to map samples from uniform distributions to the probability density of the illumination, also providing their corresponding probabilities; and an implicit neural representation which compresses the environment map into an efficient differentiable function. The computation time of environment samples with NEnv is two orders of magnitude less than with traditional methods. NEnv makes no assumptions regarding the content (i.e. natural illumination), thus achieving higher generality than previous learning-based approaches. We share our implementation and a diverse dataset of trained neural environment maps, which can be easily integrated into existing rendering engines.
  • Thumbnail Image
    Item
    SoftSMPL: Data-driven Modeling of Nonlinear Soft-tissue Dynamics for Parametric Humans
    (The Eurographics Association and John Wiley & Sons Ltd., 2020) Santesteban, Igor; Garces, Elena; Otaduy, Miguel A.; Casas, Dan; Panozzo, Daniele and Assarsson, Ulf
    We present SoftSMPL, a learning-based method to model realistic soft-tissue dynamics as a function of body shape and motion. Datasets to learn such task are scarce and expensive to generate, which makes training models prone to overfitting. At the core of our method there are three key contributions that enable us to model highly realistic dynamics and better generalization capabilities than state-of-the-art methods, while training on the same data. First, a novel motion descriptor that disentangles the standard pose representation by removing subject-specific features; second, a neural-network-based recurrent regressor that generalizes to unseen shapes and motions; and third, a highly efficient nonlinear deformation subspace capable of representing soft-tissue deformations of arbitrary shapes. We demonstrate qualitative and quantitative improvements over existing methods and, additionally, we show the robustness of our method on a variety of motion capture databases.