Automating garment assembly from sewing patterns remains a significant challenge due to the lack of standardized annotation protocols and the frequent absence of semantic cues. Existing methods often rely on panel labels or handcrafted heuristics, which limit their applicability to real-world, non-conforming patterns. We present AutoSew, a fully automatic, geometry-based approach for predicting stitch correspondences directly from 2D pattern contours. AutoSew formulates the problem as a graph matching task, leveraging a Graph Neural Network to capture local and global geometric context, and employing a differentiable optimal transport solver to infer stitching relationships—including multi-edge connections. To support this task, we update the GarmentCodeData dataset modifying over 18k patterns with realistic multi-edge annotations, reflecting industrial assembly scenarios. AutoSew achieves 96% F1-score and successfully assembles 73.3% of test garments without error, outperforming existing methods while relying solely on geometric input. Our results demonstrate that geometry alone can robustly guide stitching prediction, enabling scalable garment assembly without manual input. Our dataset and code are available online.1

Flatbed scanners have emerged as promising devices for high-resolution, single-image material capture. However, existing approaches assume very specific conditions, such as uniform diffuse illumination, which are only available in certain high-end devices, hindering their scalability and cost. In contrast, in this work, we introduce a method inspired by intrinsic image decomposition, which accurately removes both shading and specularity, effectively allowing captures with any flatbed scanner. Further, we extend previous work on single-image material reflectance capture with the estimation of opacity and transmittance, critical components of full material appearance (SVBSDF), improving the results for any material captured with a flatbed scanner, at a very high resolution and accuracy. • We introduce a generative model for digitizing materials using any flatbed scanners, capable of removing undesirable shading and specular highlights. • We expand the realism of digital replicas of material by including opacity and transmittance in the material model, both of which are key attributes for thin-layer materials like fabrics. • We provide an extensive and thorough experimentation using image-based and render aware metrics. Our results show that our method works with a wide range of scanning devices.

Selection is the first step in many image editing processes, enabling faster and simpler modifications of all pixels sharing a common modality. In this work, we present a method for material selection in images, robust to lighting and reflectance variations, which can be used for downstream editing tasks. We rely on vision transformer (ViT) models and leverage their features for selection, proposing a multi-resolution processing strategy that yields finer and more stable selection results than prior methods. Furthermore, we enable selection at two levels: texture and subtexture, leveraging a new two-level material selection (DuMaS) dataset which includes dense annotations for over 800,000 synthetic images, both on the texture and subtexture levels.

Advances in NERFs have allowed for 3D scene reconstructions and novel view synthesis. Yet, efficiently editing these representations while retaining photo realism is an emerging challenge. Recent methods face three primary limitations: they're slow for interactive use, lack precision at object boundaries, and struggle to ensure multi-view consistency. We introduce IReNe to address these limitations, enabling swift, near real-time color editing in NeRF. Leveraging a pre-trained NeRF model and a single training image with user-applied color edits, IReNe swiftly adjusts network parameters in seconds. This adjustment allows the model to generate new scene views, accurately representing the color changes from the training image while also controlling object boundaries and view-specific effects. Object boundary control is achieved by integrating a trainable segmentation module into the model. The process gains efficiency by retraining only the weights of the last network layer. We observed that neurons in this layer can be classified into those responsible for view-dependent appearance and those contributing to diffuse appearance. We introduce an automated classification approach to identify these neuron types and exclusively fine-tune the weights of the diffuse neurons. This further accelerates training and ensures consistent color edits across different views. A thorough validation on a new dataset, with edited object colors, shows significant quantitative and qualitative advancements over competitors, accelerating speeds by 5x to 500x.

We introduce TexTile, a novel differentiable metric to quantify the degree upon which a texture image can be concatenated with itself without introducing repeating artifacts (i.e., the tileability). Existing methods for tileable texture synthesis focus on general texture quality, but lack explicit analysis of the intrinsic repeatability properties of a texture. In contrast, our TexTile metric effectively evaluates the tileable properties of a texture, opening the door to more in-formed synthesis and analysis of tileable textures. Under the hood, TexTile is formulated as a binary classifier carefully built from a large dataset of textures of different styles, semantics, regularities, and human annotations. Key to our method is a set of architectural modifications to baseline pretrain image classifiers to overcome their shortcomings at measuring tileability, along with a custom data augmen-tation and training regime aimed at increasing robustness and accuracy. We demonstrate that TexTile can be plugged into different state-of-the-art texture synthesis methods, in-cluding diffusion-based strategies, and generate tileable textures while keeping or even improving the overall texture quality. Furthermore, we show that TexTile can objectively evaluate any tileable texture synthesis method, whereas the current mix of existing metrics produces uncorrelated scores which heavily hinders progress in the field.

We present a method for estimating the intrinsic components of a dynamic scene captured with multi-view video sequences. Unlike previous work focused either on static scenes or single view videos, our method simultaneously addresses the challenges of dealing with the extra computational complexity given by the dynamic motion while enables novel view synthesis. Key to our method to make the output temporally consistent is to encode the temporal information in a latent embedding that leverages the redundant information of the dynamic scene. Our intrinsic components includes diffuse and specular albedo, as well as scene geometry and environment illumination. We explicitly account for light visibility, which we estimate efficiently by considering dynamic and static points separately, making the problem computationally tractable. We demonstrate the effectiveness of our approach through quantitative and qualitative experiments, showing that it outperforms the naïve per-frame decomposition approach in several real-world scenes.

Abstract Estimating fabric mechanical properties is crucial to create realistic digital twins. Existing methods typically require testing physical fabric samples with expensive devices or cumbersome capture setups. In this work, we propose a method to estimate fabric mechanics just from known manufacturer metadata such as the fabric family, the density, the composition, and the thickness. Further, to alleviate the need to know the fabric family –which might be ambiguous or unknown for nonspecialists–we propose an end‐to‐end neural method that works with planar images of the textile as input. We evaluate our methods using extensive tests that include the industry standard Cusick and demonstrate that both of them produce drapes that strongly correlate with the ground truth estimates provided by lab equipment. Our method is the first to propose such a simple capture method for mechanical properties outperforming other methods that require testing the fabric in specific setups.

We propose a learning-based method to recover normals, specularity, and roughness from a single diffuse image of a material, using microgeometry appearance as our primary cue. Previous methods that work on single images tend to produce over-smooth outputs with artifacts, operate at limited resolution, or train one model per class with little room for generalization. In contrast, in this work, we propose a novel capture approach that leverages a generative network with attention and a U-Net discriminator, which shows outstanding performance integrating global information at reduced computational complexity. We showcase the performance of our method with a real dataset of digitized textile materials and show that a commodity flatbed scanner can produce the type of diffuse illumination required as input to our method. Additionally, because the problem might be ill-posed –more than a single diffuse image might be needed to disambiguate the specular reflection– or because the training dataset is not representative enough of the real distribution, we propose a novel framework to quantify the model's confidence about its prediction at test time. Our method is the first one to deal with the problem of modeling uncertainty in material digitization, increasing the trustworthiness of the process and enabling more intelligent strategies for dataset creation, as we demonstrate with an active learning experiment.

Abstract 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.

Abstract 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.

Existing devices for measuring material appearance in spatially-varying samples are limited to a single scale, either micro or mesoscopic. This is a practical limitation when the material has a complex multi-scale structure. In this paper, we present a system and methods to digitize materials at two scales, designed to include high-resolution data in spatially-varying representations at larger scales. We design and build a hemispherical light dome able to digitize flat material samples up to 11x11cm. We estimate geometric properties, anisotropic reflectance and transmittance at the microscopic level using polarized directional lighting with a single orthogonal camera. Then, we propagate this structured information to the mesoscale, using a neural network trained with the data acquired by the device and image-to-image translation methods. To maximize the compatibility of our digitization, we leverage standard BSDF models commonly adopted in the industry. Through extensive experiments, we demonstrate the precision of our device and the quality of our digitization process using a set of challenging real-world material samples and validation scenes. Further, we demonstrate the optical resolution and potential of our device for acquiring more complex material representations by capturing microscopic attributes which affect the global appearance: we characterize the properties of textile materials such as the yarn twist or the shape of individual fly-out fibers. We also release the SEDDIDOME dataset of materials, including raw data captured by the machine and optimized parameteres.

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.

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.

Abstract 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.

Abstract 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.

Efficient hardware implementations routinely approximate mathematical functions with look-up tables, while keeping the error of the approximation under control. For a certain class of commonly occurring 1D functions, namely monotonically increasing or decreasing functions, we found that it is possible to approximate such functions by repeated application of a very low resolution 1D look-up table. There are many advantages to cascading multiple identical LUTs, including the promise of a very simple hardware design and the use of standard linear interpolation. Further, the complexity associated with unequal bin sizes can be avoided. We show that for realistic applications, including gamma correction, high dynamic range encoding and decoding curves, as well as tone mapping and inverse tone mapping applications, multiple cascaded look-up tables can reduce the approximation error by more than 50% compared to a single look-up table with the same total memory footprint.

The estimation of the optical properties of a material from RGB-images is an important but extremely ill-posed problem in Computer Graphics. While recent works have successfully approached this problem even from just a single photograph, significant simplifications of the material model are assumed, limiting the usability of such methods. The detection of complex material properties such as anisotropy or Fresnel effect remains an unsolved challenge. We propose a novel method that predicts the model parameters of an artist-friendly, physically-based BRDF, from only two low-resolution shots of the material. Thanks to a novel combination of deep neural networks in a nested architecture, we are able to handle the ambiguities given by the non-orthogonality and non-convexity of the parameter space. To train the network, we generate a novel dataset of physically-based synthetic images. We prove that our model can recover new properties like anisotropy, index of refraction and a second reflectance color, for materials that have tinted specular reflections or whose albedo changes at glancing angles.

We present a model to measure the similarity in appearance between different materials, which correlates with human similarity judgments. We first create a database of 9,000 rendered images depicting objects with varying materials, shape and illumination. We then gather data on perceived similarity from crowdsourced experiments; our analysis of over 114,840 answers suggests that indeed a shared perception of appearance similarity exists. We feed this data to a deep learning architecture with a novel loss function, which learns a feature space for materials that correlates with such perceived appearance similarity. Our evaluation shows that our model outperforms existing metrics. Last, we demonstrate several applications enabled by our metric, including appearance-based search for material suggestions, database visualization, clustering and summarization, and gamut mapping.

To enable light fields of large environments to be captured, they would have to be sparse, i.e. with a relatively large distance between views. Such sparseness, however, causes subsequent processing to be much more difficult than would be the case with dense light fields. This includes segmentation. In this paper, we address the problem of meaningful segmentation of a sparse planar light field, leading to segments that are coherent between views. In addition, uniquely our method does not make the assumption that all surfaces in the environment are perfect Lambertian reflectors, which further broadens its applicability. Our fully automatic segmentation pipeline leverages scene structure, and does not require the user to navigate through the views to fix inconsistencies. The key idea is to combine coarse estimations given by an over-segmentation of the scene into super-rays, with detailed ray-based processing. We show the merit of our algorithm by means of a novel way to perform intrinsic light field decomposition, outperforming state-of-the-art methods.

Abstract 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.

Abstract 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. The recent introduction of compressive sensing and sparse reconstruction techniques allows for the capture of single‐shot 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 convolutional sparse coding (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.

This paper presents a method for measuring the similarity in style between two pieces of vector art, independent of content. Similarity is measured by the differences between four types of features: color, shading, texture, and stroke. Feature weightings are learned from crowdsourced experiments. This perceptual similarity enables style-based search. Using our style-based search feature, we demonstrate an application that allows users to create stylistically-coherent clip art mash-ups.

We present a method to decompose a video into its intrinsic components of reflectance and shading, plus a number of related example applications in video editing such as segmentation, stylization, material editing, recolorization and color transfer. Intrinsic decomposition is an ill-posed problem, which becomes even more challenging in the case of video due to the need for temporal coherence and the potentially large memory requirements of a global approach. Additionally, user interaction should be kept to a minimum in order to ensure efficiency. We propose a probabilistic approach, formulating a Bayesian Maximum a Posteriori problem to drive the propagation of clustered reflectance values from the first frame, and defining additional constraints as priors on the reflectance and shading. We explicitly leverage temporal information in the video by building a causal-anticausal, coarse-to-fine iterative scheme, and by relying on optical flow information. We impose no restrictions on the input video, and show examples representing a varied range of difficult cases. Our method is the first one designed explicitly for video; moreover, it naturally ensures temporal consistency, and compares favorably against the state of the art in this regard.

Abstract Many high‐level image processing tasks require an estimate of the positions, directions and relative intensities of the light sources that illuminated the depicted scene. In image‐based rendering, augmented reality and computer vision, such tasks include matching image contents based on illumination, inserting rendered synthetic objects into a natural image, intrinsic images, shape from shading and image relighting. Yet, accurate and robust illumination estimation, particularly from a single image, is a highly ill‐posed problem. In this paper, we present a new method to estimate the illumination in a single image as a combination of achromatic lights with their 3D directions and relative intensities. In contrast to previous methods, we base our azimuth angle estimation on curve fitting and recursive refinement of the number of light sources. Similarly, we present a novel surface normal approximation using an osculating arc for the estimation of zenith angles. By means of a new data set of ground‐truth data and images, we demonstrate that our approach produces more robust and accurate results, and show its versatility through novel applications such as image compositing and analysis.

Abstract Decomposing an input image into its intrinsic shading and reflectance components is a long‐standing ill‐posed problem. We present a novel algorithm that requires no user strokes and works on a single image. Based on simple assumptions about its reflectance and luminance, we first find clusters of similar reflectance in the image, and build a linear system describing the connections and relations between them. Our assumptions are less restrictive than widely‐adopted Retinex‐based approaches, and can be further relaxed in conflicting situations. The resulting system is robust even in the presence of areas where our assumptions do not hold. We show a wide variety of results, including natural images, objects from the MIT dataset and texture images, along with several applications, proving the versatility of our method.